Quickest Flow

HOLY CROSS COLLEGE OF CALINAN
CALINAN, DAVAO CITY
THE EFFECTIVENESS OF QUICKEST FLOW MODEL

IN LESSENING THE EVACUATION TIME
IN HOLY CROSS COLLEGE OF CALINAN
Ong, Tian Ruy

Plana, Eloisa Marie
Ruta, Nikki Jane
Morales, Jonh Kenneth
Acebu, Reynaldo Jr.
Lozada, Alyssa Marie
October, 2017
i
THE EFFECTIVENESS OF QUICKEST FLOW MODEL

IN LESSENING THE EVACUATION TIME
IN HOLY CROSS COLLEGE OF CALINAN
______________________________________
A Research Paper
Presented to the Faculty of the Basic Education Department
of the Holy Cross College of Calinan
_______________________________________
In Partial Fulfillment of the Requirements

in Practical Research II
By:
Ong, Tian Ruy

Plana, Eloisa Marie
Ruta, Nikki Jane
Morales, Jonh Kenneth
Acebu, Reynaldo Jr.
Lozada, Alyssa Marie
October, 2017
ii
APPROVAL SHEET
In partial fulfillment of the requirements in Practical Research II, this study
entitled THE EFFECTIVENESS OF QUICKEST FLOW MODEL IN LESSENING
THE EVACUATION TIME IN HOLY CROSS COLLEGE OF CALINAN,
prepared and submitted by Tian Ruy Ong, Eloisa Marie Plana, Nikki Jane Ruta, Jonh
Kenneth Morales, Reynaldo Acebu Jr., and Alyssa Marie Lozada is hereby
recommended for oral examination, approval and acceptance.
MELINA C. GONZALES, MAST, MES

Adviser
PANEL OF EXAMINERS
Approved by the panel of examiners, after the presentation of the study with the
grade of PASSED.
ENG. MARLOU O. BIBANGCO MA. CORAZON C. SUNGA, M.A.

Panel Member Panel Member
RIZALITO H. PAGA, M.A.

Editor
Accepted in partial fulfillment of the requirements in Senior High School.
Date of Oral Examination: October 5, 2017
MA. CORAZON C. SUNGA, M.A.

B.Ed. Principal
iii
ACKNOWLEDGEMENT
This research study would not be possible without the presence of the many
significant individuals who helped in its completion. Because of this, the researchers
would like to acknowledge and express their deepest and heart-felt gratitude to the
following:
To their parents, Mr. and Mrs. Ong, Mrs. Plana, Mr. and Mrs. Morales, Mr. and
Mrs. Lozada, Mr. and Mrs. Acebu and Mr. and Mrs. Ruta, for providing them financial
assistance and moral support during the conduct of this research study;
To the School President, Sister Lorna B. Murillo, P.M., the Basic Education
Principal, Miss Ma. Corazon C. Sunga, and the Dean of College, Dr. Mary C. Deocariza
for allowing the researchers to conduct the study;
To Mr. Van Anthony Quiamco, Miss Shogar Echavez, Mrs. Jessica Rabang, Mrs.
Luz Palarca, Mrs. Catherine Divinagracia and Miss Denise Claire Parone for their
support and for sharing their time and expertise during the data gathering procedures and
implementation of the study;
To the Disaster Action Team (DAT) especially the Security Team for helping the
researchers facilitate the flow of the students during evacuation;
To Mr. Raffy C. Centeno, for the support and encouragement which helped the
researchers address different setbacks and problems encountered in the study;
To Miss Melina Gonzales, their research adviser, who provided them the basic
knowledge on doing research and for continually supporting and guiding them
throughout their study;

iv
To the panel of examiners, Miss Ma. Corazon C. Sunga and Mr. Marlou C.
Bibanco, for their suggestions and constructive comments that helped the researchers
improve their study;
To the Basic Education students and teachers, for their cooperation and
participation during the speed tests, meetings and drills conducted by the researchers;
Above all, to God Almighty Father for giving the researchers the courage,
strength, wisdom, and enlightenment to finish the study. Thank You.

v
ABSTRACT
In this research study, an earthquake evacuation strategy was developed through
mathematical and strategic approaches using a network model: the Quickest Flow Model.
Its purpose is to find the most time-efficient evacuation routing system for the PM
building in Holy Cross College of Calinan. This is to lessen the overall evacuation time,
as well as the possible number of casualties whenever an earthquake disaster occurs.
All nodes, arcs and sinks in the school were identified and translated into a
network model. Through Djkistra’s algorithm, the shortest path each source may take to
evacuate among the given set of paths were identified, considering the arc costs, arc and
sink capacities, traffic congestion and availability of arcs at each time interval. To test the
model, two sets of drills were conducted using the non-model-based evacuation plan and
the model-based evacuation plan. The statistical tool – independent t-test, was then used
to test if there is a significant difference between the time of evacuation with and without
the model. Compared to the non-model-based evacuation plan, the total evacuation time
spent by the evacuees is significantly lower by up to 45 seconds when the model-based
evacuation plan was used. Therefore, it was concluded that the proposed model-based
evacuation plan improved the time spent during evacuation. Thus, it is recommended that
the proposed strategy be employed as the school’s new evacuation plan.
Keywords: time-efficient evacuation, Quickest Flow Model, routing strategy, shortest
path algorithm, non-model-based evacuation plan

vi
TABLE OF CONTENTS
Title Page i
Approval Sheet ii
Acknowledgement iii
Abstract v
Table of Contents vi
List of Tables ix
List of Figures x
Chapter 1 INTRODUCTION
Background of the Study 1
Statement of the Problem 5
Hypotheses 6
Review of Related Literature and Studies 6
Theoretical Framework 18
Conceptual Framework 20
Significance of the Study 20
Scope and Delimitation of the Study 21
Definition of Terms 23
Chapter 2 METHODS
Research Design 25
Research Respondents 26
Research Locale 27
Data Gathering Procedure 28
Data Analysis 29
Chapter 3 RESULTS AND DISCUSSION

Presentation, Discussion and Interpretation of Data 31
vii
Chapter 4 CONCLUSION AND RECOMMENDATIONS

Conclusion 59
Recommendations 60
References 61
Appendices
Appendix 1a: Time spent using the non-model evacuation plan 66
with observations
Appendix 1b: Time spent using the model evacuation plan with 67
observations
Appendix 2: The Applicability of the mathematical model 68
based on the demands of the school’s present
egress system
Appendix 3a: Distances from Sources to Sinks 69
Appendix 3b: Area Measurements of the Arcs 70
Appendix 4: Cost of Different Arcs in Time Units 71
Appendix 5: Capacity of different arcs relative to flow size 72
Appendix 6: Capacity of different sinks based on the average 73
occupancy
Appendix 7: Distribution of Flows at each Time Interval 74
Appendix 8: Shortest Path Algorithm 75
Appendix 9: Arrangement of Sources in the Evacuation Areas 76
Appendix 10: Arc Costs and Capacities 77
Appendix 11: Calculation of Sink Capacities 78
Appendix 12: Distribution of Flow Units at Each Time Interval 79

viii
Appendix 13a: Route Assignment Based on QFP 80
Appendix 13b: Route Assignment without the Model 81
Appendix 14: Statistical Analysis in R 82
Appendix 15: Boxplot Function 83
Appendix 16a: Formulated Evacuation Plan Using the Model in 84
PM Building Classrooms
Appendix 16b: Formulated Evacuation Plan Using the Model in 87
Different Offices
Appendix 17: Earthquake Drills Using the Non-Model-Based 88
Evacuation Plan
Appendix 18: Earthquake Drills Using the Model-Based 90
Evacuation Plan
Appendix 19: Measuring of School Dimensions 91
Appendix 20: Implementation of the Model 92
Appendix 21a: Letter of Permission to the School President 93
Appendix 21b: Letters of Permission to the School Principal 94
Appendix 21c: Letter of Permission to the Dean of College 95
Appendix 21d: Excuse Letter 96
Curriculum Vitae 97
ix
LIST OF TABLES
Table 1 Time spent using the non-model evacuation plan with observations 33
Table 2 The applicability of the mathematical model based on the demands 35
of the school’s present egress system
Table 3 Costs of different arcs in time units 43
Table 4 Capacity of different arcs relative to flow size 45
Table 5 Capacity of different sinks based on the average occupancy 46
Table 6 Distribution of flows at each time interval 48
Table 7 Time spent using the model-based evacuation plan with 55
observations
Table 8 Determining the difference using t-test 57

x
LIST OF FIGURES
Figure 1 The conceptual framework of the study showing the relationship 20
between the components of evacuation planning and the Quickest
Flow Model
Figure 2 Present Non-model Based Evacuation Plan of HCCC 31
Figure 3.1 Shortest Path Algorithm of Sources 1 and 2 38
Figure 3.2 Shortest Path Algorithm of Sources 3, 4 and 5 39
Figure 3.3 Shortest Path Algorithm of Sources 6, 7, 8, 9 and 10 39
Figure 3.5 Shortest Path Algorithm of Source 13 40
Figure 3.6 Shortest Path Algorithm of Sources 14, 15 and16 41
Figure 4 Model-based Evacuation Plan 50
Figure 5.1 Assignment of Evacuees in Football Sink 51
Figure 5.2 Assignment of Evacuees in Front of Guidance offices 52
Figure 5.3 Assignment of Evacuees in Pieta Sink 53
Figure 5.4 Assignment of Evacuees in Grade School Ground Sink 53

CHAPTER 1
INTRODUCTION
This chapter presents the purpose of conducting the study, its significance, the
parameters to be considered and the hypotheses to be tested in the study. It also evaluates
existing related literatures and specifies the variables that will have to be explored.
Background of the Study
In the past decades, disasters have perceptibly increased in number. In the years
1994-2003, Asia was greatly affected by natural disasters in which approximately 70 out
of the 650 natural catastrophes recorded in 2004 were damaging earthquakes, 10 were
volcanic eruptions, and the rest were strong weather conditions. The 1990s and beyond
proved to be much calamitous, as global natural disaster trends rose with increasing
rapidity (Sena & Woldemichael, 2006). Hence, strategic evacuation response is necessary
to reduce the immediate and potential life-threatening dangers posed by such disasters
(Peeta, Sharma, & Hsu, 2011). Although many evacuation models have been proposed,
accepted, and simulated in the past years, Ronchi and Nilsson (2013) noticed that in most
cases, buildings targeting a safe evacuation often never go beyond the simplified
approach. This, however, should not be the case as preparedness is a key component of a
major incident management, which actually plays a significant role in the emergency
response (Castle, 2006). In designing an evacuation plan, the travel time movement
becomes the greatest factor. Thus, a good routing strategy will evacuate as many
evacuees as possible within a limited time (Ng & Chow, 2006; Qui & Jin, 2008).
2
There are many reasons that cause emergency people movement in a building
such as fire, earthquake, typhoons and civil disobedience. Traditionally, dealing with
emergency evacuation is by code compliance. Codes consider availability of egress
routes but do not consider the effects of potential blockage of ways out on time
evacuation. Even though various egress designs can be code compliance, they will still
perform differently in emergency evacuations. Some lead to a shorter evacuation time
while others prevent occupants to escape efficiently. Therefore, different kinds of layouts
can be compared in their evacuation performance. But regardless of the layout, an
appropriate configuration of plan design can provide a better evacuation result.
According to Gwynne and Galea (1998) as cited by Castle (2006), to perfectly assess the
potential evacuation efficiency of a building or an enclosure, it is necessary to consider
four aspects, which are configuration, environment, behavior and procedures.
When it comes to emergency preparedness, Japan has greatly progressed. The
country’s disaster management system has undergone tremendous advancement
throughout the past 5-6 decades (Nazarov, 2011). According to Yamada (1996) as cited
by Xie, Lin and Waller (2010), under the Natural Disaster Prevention and Relief Law,
every Japanese city has its own emergency evacuation plan in preparation for such major
disasters as earthquakes, tidal waves, etc. The plan includes places of refuge and the
allocation of all the residents to the refuge areas. In spite of this, the magnitude 9.0
earthquake incident at the coast of North-eastern Japan on March 11, 2011 showed the
devastating effect of the disaster destroying majority of high-rise buildings killing about
15, 891 people with 2,500 missing cases. The ground displacement during the earthquake
almost instantly moved the lower levels of the high-rise buildings in Japan, but it takes a
3
moment for the shaking to reach the upper floors. This created a dangerous seismic echo
in the structures, wherein the upper stories rushed to catch up with the motion of the
floors below. In just moments of severe shaking, it resulted to structural failure and
complete collapse of the high-rise buildings.
Being in the Pacific Ring of Fire just like Japan, the Philippine islands have
experienced up to twenty earthquakes a day (Group, 2012), although most are too deep or
too small to be felt. In 2008, the area around Ta’al volcano alone recorded ten volcanic
earthquakes in just one day. According to Sieh (2007), Asia has a geological record that
goes back 1,000 years. It shows the region (South-East Asia) being hit by major
earthquakes every 200 - 300 years. The last cluster of powerful quakes happened about
200 years ago which statistically means that South-East Asia is entering a new cluster or
new set of powerful earthquakes. In connection, in the year 1990, a 7.8 magnitude
earthquake shook Northern and Central Luzon with its epicenter recorded in Nueva Ecija
which lasted for about a minute as recorded by Rappler Philippines. Among the hardest
hit areas was Baguio City in which several structures collapsed, burying people alive.
Some establishments were completely destroyed including hotels like the 12-storey Hyatt
Terraces Plaza, 5-storey Nevada Hotel, 6-storey Baguio Hilltop Hotel, 4-storey Baguio
Park Hotel and FRB hotel which mainly suggests the susceptibility of high-rise buildings
to severe ground shaking. Because of this, building evacuation becomes a critical factor
in emergency preparedness (Lu, 2011). Since then, public and private sector planners
take very different approaches to planning for catastrophic events, so it will be instructive
to examine earthquake risk as an emergency manager would through the four phases of
emergency management - mitigation, preparedness, response and recovery.

4
Experts suggest that many schools fail to formulate policies and guidelines
regarding preparedness in terms of evacuation plans, designated evacuation area, and
safety awareness. According to UNESCO (2012), school safety is highly necessary as
children are among the most vulnerable group. In fact, educational places such as schools
are among the most important buildings exposed to serious damage and loss of life from
earthquakes (Hosseini & Izadkhah, 2006). When the 7.2 earthquake hit Sulu on January
10, 2017, the tremors were felt even in Davao City. Students and employees from
different establishments evacuated to open grounds (Revita, 2017). Because of the
earthquake disaster risk management planning, a “safety culture” was developed in which
people became aware of the hazards that may confront them and the knowledge with
which they can protect themselves (Hosseini & Izadkhah, 2006). The recent earthquake
in Davao has shown that the knowledge from regular drills helped people in reacting
during a disaster.
At present, the Holy Cross College of Calinan (HCCC) has 3 high-rise buildings
which include a 5-storey college building, 4-storey PM building and 2-storey OLSW
building that require an efficient evacuation plan considering their susceptibility to severe
ground shakings. Although the institution periodically conducts evacuation drills, the
egress system used was not based upon a model. To address such problem, this study
aims to provide a mathematical model-based evacuation plan to decrease the overall
evacuation time by lessening the congestion occurring along the arcs and by proper route
assignment. By using a model-based plan, it will give better insights and orientations for
solving real-life complex systems such as the maximization of flows and minimization of
5
time for evacuation (Amirgaliyeva, 2012) which would in turn lessen the possibility of
having casualties whenever an earthquake may occur.
Statement of the Problem
The purpose of this study is to develop a new strategic evacuation plan in Holy
Cross College of Calinan using a mathematical model, the Quickest Flow. Specifically, it
aims to answer the following questions:
Phase 1: Before the Implementation of the Quickest Flow Model
1. What is the present egress system of Holy Cross College of Calinan?
2. What is the average time spent during evacuation using the non-model-based
evacuation plan?
3. Is the Quickest Flow Model applicable for planning an evacuation strategy in
Holy Cross College of Calinan?
4. What procedures should be employed to improve egress efficiency?
Phase 2: After the Implementation of the Quickest Flow Model
1. What is the average time spent during evacuation using the model-based
evacuation plan?
2. Is there a significant difference in the time spent between the non-model and
model-based evacuation plan?

6
Hypotheses
Ha: There is a significant difference in the time spent of evacuation between the non-
model and model-based evacuation plan.
Ho: There is no significant difference in the time spent of evacuation between the
non-model and model-based evacuation plan.
Review of the Related Literature
Evacuation is defined as the removal of lives and/or property from the disaster
zone to the safety zone as quickly as possible (Khadka, 2016). It is an intuitive and
practically effective emergency rescue measure that has long been used and is expected
to be enhanced to protect human population against hazardous situations caused by
natural and man-made disasters.
The viewpoint of modelling and improving an evacuation process as a special
transportation system has been recognized since late 1990’s, and apparently stimulated
after two major hurricanes, Hurricanes Georges in 1998 and Floyd in 1999. The lessons
learned from the two mass evacuation instances indicated that many transportation and
traffic related efficiency and safety issues would be frequently encountered yet not
satisfactorily addressed at the current state of evacuation practice (Xie, 2008). Further,
when Hurricane Andrew was approaching Florida and Louisiana in 1992, the affected
population was just simply asked to leave the area as soon as possible. This caused
tremendous traffic congestion on highways and led to great confusion and chaos (Lu,
Huang & Shekhar, 2008). More recently, Hurricane Katrina, the third worst hurricane in
U.S. history, struck the Golf Coast in August 2005, resulting more than 1,800 deaths and
7
an estimated $81.2 billion loss of property. The evacuation prior to and just after this
hurricane is widely regarded as a failure of emergency response and management.
Several post-disaster investigations have cited evacuation failures as a major contributor
to the death toll in the city of New Orleans. In order to manage such emergencies more
effectively, decision makers may benefit from having in-place evacuation plans for
scenarios which are most likely to happen, even though real-time design and re-
evaluation of evacuation plans may be required after the disaster strikes (Alexander,
2002, as cited by Stepanov & Smith, 2008).
In recent years, catastrophic disasters by massive earthquakes are increasing in the
world making disaster management more required than ever. An example of this is the
Tohoku-Pacific Ocean Earthquake incident that happened in Japan on March 11, 2011
which caused a tsunami that brought serious damage to certain areas. Although disaster
prevention in civil and architectural engineering fields of Japan had been considered
mainly from physical aspects, previously, it becomes clear that it is difficult to prevent
large tsunami physically. Therefore, the disaster prevention from non-physical aspect,
such as city planning and evacuation planning, is considered to become more important
from now on (Katoh, Takizawa, & Inoue, 2012).
Evacuation Plan
Emergency evacuation plans assign evacuees to fixed routes or destinations even
before disasters. An evacuation plan defines optimal evacuation policies for the
population from areas under risk and uncertainty (Stepanov & Smith, 2008). According
to Xie (2008), a well-defined and manageable plan based on network models is one of the
prerequisites for successful implementation of a large-scale evacuation. The purpose of

8
an evacuation plan is to maximize the utilization of an existing transportation system for
evacuation of a threatened population and hence to minimize the exposure and potential
fatalities of the population from the impending or occurring disaster. However,
preventive measures are to be taken into account during the construction of the
evacuation plan (Khadka, 2016). These preventive measures can take many forms
including the construction of evacuation shelters, implementation of emergency
communication system and installation of warning devices.
Evacuation Planning
Evacuation planning, on the other hand, refers to the formulation of a plan to be
used for evacuation purposes. It includes the estimation of the evacuation time,
propagation of the disaster, the potential risk and location of the safety zone and the
reorganization of the traffic routes from the disaster zone to the safety zone. According to
Xie (2008), it is an important component of emergency preparedness in urban and rural
areas in which the number and location of rescue facilities or areas is an important aspect
of this planning, as is the identification of primary and secondary evacuation routes for
the residents to take.
As one of the aspects of disaster mitigation and evacuation planning, the planners
must find or provide an effective tactic and operation strategy managing the traffic and
transportation needs during an evacuation. The development of such strategies is a
difficult task that may be greatly aided through involving mathematical traffic modeling
with optimization insights (Zheng, Chiu, Mirchandani, & Hickman, 2010). An evacuation
problem planning was formalized by using the idea of network flow, and has proposed
9
the evacuation model to find the fastest completion time of evacuees. Minimizing the
evacuation time by which all evacuees completed the evacuation is the criterion for the
evacuation planning (Takizawa, Inoue, & Katoh, 2011).
The Quickest Flow Model
The overall evacuation performance could be better assessed by evaluating the
number of evacuees having left the disaster area over time. Such an evacuation process
requires a simulation model, or in other words, in these cases, an integrated optimization-
simulation approach is highly desirable (Xie, 2008).
In the context of evacuation modeling, dynamic network flow theory can be used
to provide lower bounds on evacuation times and optimal evacuation routes. It aims to
send a maximum amount of flow from a source to a sink within a given time bound
(Pyakurel & Dhamala, 2014). An optimal dynamic network flow can be computed
efficiently and yields a best achievable evacuation plan, provided all evacuees follow the
routes corresponding to this flow (Hamacher, Leiner, & Ruzika, 2011). According to
Hooks and Patterson (2004) as cited by Baumann and Skutella (2009), the importance
and wide applicability of network flow problems has long been recognized. Dynamic
network flow problems, in particular, have been used to model numerous real world
phenomena arising in applications within almost all industries. Such applications include
production-distribution systems, fleet management, evacuation and communications.
One of the well-studied dynamic network flow problem in the context of
evacuation modeling is the quickest flow problem (QFP): given a dynamic network and a
number b of flow units, find the minimal time T needed to route all flow units from the
source s to the target t. It aims to send a given flow amount from the source to the sink in
10
the shortest possible time. The resulting minimal time T therefore gives a lower bound on
the minimal evacuation time (Hamacher, Leiner & Ruzika, 2011). In addition, evacuation
for pedestrians from buildings based on dynamic network flow has also been extensively
studied. The problem maximizes a number of evacuees and evacuation time. The disaster
zone may have multiple nodes. Another variant with time dependent flow uses tree
structure network to evacuate from common buildings. Moreover, time dependent case
updates the evacuation routes in time period which minimizes the overall evacuation time
(Khadka, 2016).
According to Xie (2008) and Khadka (2016), many factors should be considered
when formulating an evacuation plan. Thus in this study, the following factors are
considered:
a. Traffic Congestion
During disasters it is expected that mass evacuation is performed immediately in
order to ensure safety and avoid life-threatening dangers resulting in many loss of lives
and properties. One of the major setbacks of evacuation often starts from a surge of
traffic demand which might exceed the capacity of an existing roadway system that
results in congestion or even gridlock of the network (Kalafatas & Peeta, 2009). This is in
consonance to the study of Xie (2008) which stated that evacuation by nature is a
transportation activity and with transportation, there is always traffic involved. One of the
inherited properties of the problem is that the traffic load is where the evacuation network
performance depends on. The over-utilization of a particular road link can block traffic
through the link completely. This blocking will affect all upstream links which causes the
11
significant congestion through the evacuation network. Thus according to Cruz, Talebi
and Smith (2005) as cited by Stepanov and Smith (2008), the improvement of the
network performance depends on the efficient management of the demand-supply
problem of the evacuation network
According to Zimmerman (2010), the efficient and expeditions flow of evacuation
traffic is the most critical element in a successful evacuation and the most challenging
mission, especially in a no-notice environment. The variability of the traffic management
plan employed during an evacuation will directly influence the safety and comfort of the
evacuees. Compared with the already complicated traffic operations problem under
normal traffic conditions, operations during emergency evacuation can be much more
difficult. Significantly, heavier flows throughout the network result in a rapid rise in
demand. Excessive queues and delays are expected under such scenarios, as the demand
volume far exceeds the capacity of the network. Improper evacuation plans will cause
some areas to be more congested than others, reducing the opportunity for escape from
these areas.
b. Traffic Route Assignment
In recent years, devising an optimal evacuation plan has been advanced from
seeking the optimal routing and scheduling scheme in a given network and spreading
evacuees to follow the corresponding system-optimal evacuation order, to promoting the
optimal flow pattern by physically manipulating the network configuration. In these
network planning models, the decision variables not only reside in demand routing and
12
scheduling, but also include network supply properties, such as contraflow configuration
and intersection control which need an appropriate route assignment (Xie, 2008).
Traffic route assignment methods have been one of the most widely used
approaches to model traffic flow over road networks since the first mathematical
formulation of static traffic assignment problem was proposed by Beckmann (1956) as
cited by Chang, Elnashai, Spencer, Song and Ouyang (2010). The basic idea is to
recognize the traffic routing from the disaster zone to the safety zone. The disaster zone
and/or safety zone may have multiple nodes. The disaster zone may be buildings,
vehicles, stadiums, cities or a region. The main objective of traffic planning is prior
selection of scalable traffic routes for evacuees within a desired time (Khadka, 2016).
When planning an evacuation traffic route assignment, the areas, destinations and
evacuation transportation networks must be properly defined (Kerbache & Smith, 2000 as
cited by Stepanov, et. al., 2008).
c. Human Behavior Under Emergency Conditions
Evacuee behavior is a complex phenomenon that needs to be understood
(Stepanov & Smith, 2008) since it causes most of the clogs during evacuations. Valid
assumptions on human behavior during emergencies such as panic, herding, congestion,
and ignorance of guidance help design a fixed system for safe and efficient evacuation.
Although, it is also a challenge as behavior may vary to each person and it is plagued
with uncertainties especially in times of panic (Caunhye, Nie, & Pokharel, 2011). The
social and familiarity factors of evacuees have also been noticed by Shen (2005) as cited
by Chizari, et. al. (2013), where mostly the effect of proposed model is on the walking
13
speed and the route selection is still based on shortest path to exit. On the way to exit,
evacuees, who are in panic mode, may push each other and this will reduce the
evacuation process. More often than not, the behavioral pattern is consistent across
different emergency conditions (Pel, Bliemer, & Hoogendoorn, 2011).
In addition, since evacuation is a one-time transportation activity, evacuees may
not have sufficient experience and adequate information to make proper routing and other
travel choices. Their travel decisions and other evacuation related judgments are highly
dependent on their own perceptions of the risk, as well as the information transferred
from their social networks and the authority. According to Xie (2008), due to the
unpredictability and suddenness of disasters, their perceived information is often
inaccurate and incomplete. As a result, travel behavior of evacuees is uncertain and
disordered compared to their ordinary travel activities. Nevertheless, most studies
suggested that human behavior is still a necessary component when building an
evacuation plan (Castle, 2006; Pel et. al., 2011; Pelechano & Badler, 2006; Sagun,
Anumba, & Bouchlaghem, 2013).
In the study, the evacuee behavior will be divided into two categories: behavior
under emergency conditions and the behavior during evacuation drills. However, since
the Quickest Flow Model cannot take drill behavior such as the laughing, shouting, and
bystanding into account, assumptions have been formulated instead.
d. Identification of Refuge Centers
Emergency shelter is the temporary living place for people to avoid the danger of
natural disasters such as earthquake, fire, flood and so on. Based on scientific planning
14
and standard management, the emergency shelter could be used to supply the basic
substinence requirements for people in dangers (Yang, Han, Zhang & Song, n.d.).
In addition, an evacuation area should be away from the lots that are prone to
earthquake fault zone or where there are dangerous and adverse hazards. It should be flat,
open and should not be flooded by river or other secondary disasters. Effective refuge
area should exclude water area, low shrubs, areas covered by trees, non-refuge buildings
and area that slope is greater than 15 degrees (Wang, Wang, Liu, & Chu, 2013) and most
especially it should have sufficient reserve space to accommodate more people during
evacuation. For evacuation purpose, the design must designate space of 1m2 per person
(Bappenas, 2005).
The evacuation area should be established and evacuees should be notified of
their locations in advance to reduce the risks and costs (Karnema, Nohadeseh, Samira &
Asghari, 2014). The concept of relaxing the constraint of assigning evacuees to
prespecified destinations will most likely improve evacuation efficiency (Yuan, Han,
Chin & Hwang, 2006). According to Saadatseresht, Mansourian, and Taleai (2009), a
central challenge in developing an evacuation plan is determining the distribution of
evacuees. At this step of evacuation planning process, the destinations should be
identified (Lindell & Prater, 2007).
e. Building Evaluation
In recent years, super high-rise buildings which are greater than 500 meters are
developing very quickly and became an important frontier of civil engineering. However,
the consequent occurrence of natural disasters such as earthquakes raises awareness

15
regarding the safety of building designs (Sagun, et. al., 2013). Because of this, the
collapse resistance of super high-rise buildings subjected to extremely strong earthquake
is a critical problem that must be intensively studied (Lu, 2011). This was supported by
the study of Burtles and Noakes-Fry (2016) stating that the building structure can also be
a challenge during emergency evacuation especially for multi-storey buildings. The
higher the building, the more complex evacuation becomes. The sense of awareness on
this issue was raised by the World Trade Center terrorist attack of 9/11 (Ronchi &
Nilsson, 2013). When the terrorists attacked WTC, it took as long as six to eight hours for
the occupants to exit the building. Furthermore, it showed the importance of providing
robust means of egress (Hall, 2011) by enabling efficient evacuation (Sagun et. al., 2013).
The evacuation process in high-rise building is substantially affected by the
characteristics of vertical egress components. Modern egress design should take into
account several variables, including the change of occupant demographics (Spearpoint &
Maclennan, 2012), occupant behaviors (Nilsson & Jonsson, 2011), the advance in the
technologies that lead to the design of extremely complex buildings and the subsequent
increase in the building height. Because of this, evacuation procedures in buildings have
long been studied in many aspects from architecture, engineering, computer modeling,
etc (Chizari, et. al., 2013). Despite the numerous efforts, evacuating a building keeps
getting complicated as time goes and at present, mathematical models and simulations
help create a fast and safe evacuation process.

16
Limitation of the Optimization Model
Despite the ease of implementation and less computational research required, the
main weaknesses of this optimization-based evacuation planning models are quite
difficult to address. According to Xie (2008), there are two factors that are hard to cope in
the evacuation models. First is that the central emergency management authority is
implicitly assumed to regulate the whole evacuation process while the individual
behaviors may not be appropriately taken into account and second, in many cases, a
proper relationship between traffic and travel time is not incorporated into the estimation
of the spatial distribution of evacuating flow over the network.
Hooks and Patterson (2004) as cited by Baumann and Skutella (2009) also stated
that dynamic network flow problems have been traditionally considered in a purely static
environment, i.e. the attributes of the network, including arc traversal times (or costs), arc
and node capacities, and the availability of supply, are time-invariant. For many of the
aforementioned applications, such a static representation may be inadequate. Thus, to
more realistically model such problems, the methodology must recognize the inherent
time-varying nature not only of flow, but also of the network attributes. While standard
network flows are useful to model a variety of optimization problems, they fail to capture
a crucial element of many routing problems: routing occurs over time.
In the study of Xie (2008), evacuation planning may deal with both short-term
and long-term issues. Defining the short-term evacuation plan, it needs to be enacted
quickly responding to an emergency in a predicted hazardous event. The long-term
evacuation plan, on the other hand, is generally proposed for a potential emergency area
17
in which some natural disaster may have frequently occurred in history and is expected to
occur again in a foreseeable future.
Statistical Analysis for Comparison of Data
There are many different standard tests available for comparing two distributions.
A common way to test that the difference between two means is non-random is the
independent t-test (Hammer, Harper & Ryan, 2008). However, the test should only be
used if populations are close to normally distributed. Also the t test compares the means
of two data sets, the F test compares their variances (Hammer et. al., 2008; J.S. Yuan,
Reed, Chen & Stuart, 2006).
Moreover, the data sets must not contain any strong outlier (Moore, 2004). To
identify such extreme data, representations such as the boxplot is frequently used
(Dawson, 2011). It essentially condenses the entire sample to a five-number summary,
plus locations of outliers as defined by extreme distance outside the range of the middle
half of the data (Larget, 2014). For small sample sizes, fewer number of extreme data is
expected. Nevertheless, with the small sample, results may unusually have narrow fences.
When this happens, it is more probable that data will be flagged as outliers.
The review above further elaborated this research by reviewing previous studies.
As a whole, the development of a strategic evacuation plan considers more than just the
distance and time of egress. Factors such as traffic, building structure, and behavior must
also be considered. Nakanishi, Kim, Ulusoy, and Bata (2003) as cited by Castle (2006)
stated that robust incident preparedness plans should incorporate the results of available
models. These are useful in estimating the response strategies and evacuation times.
Nevertheless, there will always be uncertainties when it comes to emergency evacuations.

18
These uncertainties should also be understood as it affects the efficiency of the
evacuation.
However, while reviewing previous studies, it was found that there was no
previous research regarding an application of an evacuation model conducted in the
school under study.
Theoretical Framework
Networks can be found anywhere in very different scientific disciplines such as
economics, organizational studies, social sciences, biology, logistics and others. In any
type of disciplines, networks are always the same in meaning: to represent graphs for
real-world objects in which nodes represent entities of the system and edges represent the
relationships among them (Steen, 2010).
Steen (2010) said that the concept of the network theory is essentially used in
describing phenomena from various fields including communication infrastructures,
drawing maps, scheduling tasks, social structures, and many other sciences. This theory
provides a set of techniques for analyzing graphs, and with complex systems, it also
analyzes structure in a system of interacting agents represented as network. In order to
describe the network structures representing real-world system, it is a need to find an
awful amount of information involving number of nodes and links, diameter, clustering,
presence of communities, and other properties (Estrada, 2013).
Constructing a strategic evacuation plan involves the use of nodes and links thus
making the network theory a critical concept in the study. Furthermore, the mathematical
model, quickest flow, is an example of a linear-programming model under the network

19
flow. This actually deals with optimization problems including shortest path, maximum
flow, and minimum cost flow problems (Ahuja, Magnanti, & Orlin, 1993).
In addition, a prime tool for modeling and designing automated decision-making
processes in interactive environments has been developed into a theory. The game theory
is a mathematical method for analyzing calculated circumstances, such as in games,
where a person's success is based upon the choices of others (Lo, Huang, Wang, & Yuen,
2006). According to Li, Zhu, Li, Wu, and Zhang (2015), an improved game theory now
finds the global minimization for the evacuation time. It helps obtain an optimum agent
distribution with estimation of the degree of risk of a route to manage the routing
selection problem and the congestion conflict. During the past decade, game theory has
been developed in the areas of competitive behaviors and the rationality problem of
evacuees. However, the degree of congestion of the emergency exit, the average
congestion degree of the evacuation route, and the maximum flow rate of exit are still
crucial factors that affect evacuation clearance time (Li, et. al., 2015).
20
Conceptual Framework
Figure 1: The conceptual framework of the study showing the relationship between
the components of evacuation planning and the Quickest Flow Model.
The Quickest Flow Model’s aim is to find the most time-efficient routing system
for evacuation. There are two main factors affecting the time of evacuation: distance from
the source to the sink (evacuation area) which is affected by the layout or design of the
building under study and also traffic congestion. The congestion is affected by the
availability of exit routes and the crowd behavior during emergency evacuation such as
panic, herding, congestion and ignorance. However, the behavior of evacuees during
emergency conditions is not considered in the study since it is not included as a variable
in the Quickest Flow Model.
Significance of the Study
A time efficient evacuation plan will be a significant endeavor for Holy Cross
College of Calinan since it promotes safety in the school. Due to the increased number of
21
students, the previous practiced evacuation plan can now be considered obsolete as it was
designed for a lesser number of evacuees. The strategic model-based evacuation plan will
help lessen the overall evacuation time by considering the necessary factors like traffic
generation and management and choice of path towards the evacuation areas. The
possible decrease in the number of casualties during earthquake calamities is one benefit
that students, teachers and administrators experience. Also, the institution can be helped
by the results of the study considering that the present evacuation plan does not follow a
network-based strategy which is a must in ensuring safety in the school. Moreover, this
study will be beneficial beyond the school’s range through inspiring and recommending
other institutions as well as the bigger community to formulate or improve their
evacuation strategy. The results of the study can also be used by the future researchers as
a guide in conducting their own studies regarding efficient evacuation planning.
Scope and Delimitation
The study focuses on formulating an efficient evacuation plan using the Quickest
Flow Model in Holy Cross College of Calinan. However, it will only be applicable to
earthquake disasters needing an evacuation. Likewise, it will not consider other factors
such as the demographic profile, the attitudes, and the level of maturity of the students
that may affect the evacuation time during emergency conditions. It will not consider the
drill behavior exhibited by the students while evacuating. Lastly, the results of this study
test only the efficiency of the model as applied to the current building floor plan of Holy
Cross College of Calinan and should not be used to measure the overall efficiency of the
network model.
22
The network model itself has its limitations which is why different assumptions
are made and considered. In this model, it is assumed through the use of arithmetic mean
that a uniform population of 45 students will occupy every classroom in the PM building.
It is also assumed that there are no debris/blockages along the exit routes that could
prevent the evacuation. The schedule is assumed to be regular which means students are
in their classrooms and that the computer laboratories are occupied. The evacuees are
assumed to proceed directly to the sink without delays or counter flows and that the
designated refuge centers are intact during evacuation.
However, during the evacuation simulation, evacuees may not perform the drill as
correctly as expected. In attempt to imitate a no-notice environment during evacuation,
the drills conducted were unannounced. Due to this, the occupancy of the laboratories
and classrooms was not certain. Also, in any given dataset, outliers may occur. These
may be caused by out-of-scope and unsuspected circumstances such as the location of the
classes during drills.
The study will only be conducted in the PM building since the number and the
dimension of the exits, the location of the sinks, and the population expected to pass a
route are not equally distributed. The school's other buildings like the Rivier, Our Lady
Seat of Wisdom, and the elementary school buildings only have one or two exits and a
single possible sink thus do not need any further optimization.

23
Definition of Terms
This section involves definitions of the unfamiliar terms present and mentioned in
the research paper that are relevant to the study.
Egress System- It is the act of going out from the classrooms and from the buildings
through exit routes in Holy Cross College of Calinan.
Evacuation- This is a temporary but rapid transferring of students and staff from a
disaster area to the safety area as a precautionary measure.
Traffic Generation- This happens when students are divided to generate traffic during
evacuations. It avoids congestion on nodes and arcs.
Algorithm- These are steps that are followed to solve a mathematical problem or
complete a process.
Quickest Flow Model- It is the model used in the study to improve the evacuation
strategy. It is composed of nodes, arcs, and sinks which are used
to map out the evacuation activity.
Network- It is a set of sources, arc, nodes and sinks.
Network nodes- These are the sources, sink, and other areas wherein traffic may occur.
These are utilized in the model and are crucial factors in formulating an
evacuation plan.
Sources- These are the classrooms in the PM Building of Holy Cross College of Calinan
where people or students originate.

24
Sinks- These are the identified destinations of evacuees during an evacuation also
known as evacuation area.
Arcs- These are the identified pathways that connect the nodes in an evacuation plan.
Flow- It is the movement of students traversing an arc or node coming from a source.
Cost- It is the time it takes to evacuate or traverse an arc measured in time units.
Time unit- It is a unit of measurement for time used in the study in which one time unit
is equal to 10 seconds.
Model-based evacuation plan- It is the proposed evacuation plan in the study following
the Quickest Flow Model.
Non-model-based evacuation plan- It is the previous evacuation plan of the school not
following a model.
CHAPTER 2
METHODS
This chapter presents the design, respondents, locale, and the methods of the
study. It considers the nature of the study and discusses the procedures the researchers
were conducting. Lastly, it explains how the data collected were analyzed.
Research Design
The study utilizes the operational method of research or systemic operational
design (SOD) which is an application of system theory to operational art. It attempts to
rationalize complexity through systemic logic and translates strategic direction and policy
into operational level designs. Moreover, it focuses upon the relationships between
entities within a system to develop a rationale for systemic behaviors that accounts for the
logic of the system thus, facilitates a cycle of design and planning (Schmitt, n.d.).
Relatively, the design is appropriate in the study because it allows the conception
and construction of a framework that can underpin a major operation such as evacuation
operation and its subsequent execution to the institution under study. The operational
design deals with the method by which the procedure is specified in the sample thus
observational and statistical designs can be conducted. Since the study focused on linear
programming, network flow programming and formulation of optimization model, then
the operational design is the most suitable to use.
The purpose of employing the operational method in the study is to come up with
a time-efficient evacuation strategy for earthquake emergencies that can be applied in the
current building floor plan of Holy Cross College of Calinan. Finding a solution for time
26
minimization involves operational design since the method attempts to apply
mathematical, logical and analytical techniques to the solution of cost minimization or of
profit maximization problems (Kothari, 2004).
The study also utilizes Quasi- Experimental research design in comparing the
time of the evacuation with and without the model. Like any other experimental designs,
it tests causal hypothesis but does not randomly assigns respondents into groups. It is
designed around an intervention which estimates the size of intervention effect on some
outcome and to test whether it differs significantly from no effect (Wiley & Sons, 2005).
The subjects are exposed to the intervention of interest then measures the outcomes from
the subjects observed. Moreover, it ascertains the effects of the independent variable
through observed changes in the dependent variable. Thus, the design is appropriate in
the study since it aims to measure whether there is a significant difference before and
after the implementation of the model-based plan. Also, it does not randomly assign the
respondents into groups.
Research Respondents
The study focused on the population occupying the college, high school and grade
school building of Holy Cross College of Calinan. Further, it utilized multistage sampling
technique to determine the average speed of the respondents during evacuation.
Multistage sampling refers to the method in which the population is divided into stages
using smaller and smaller sampling units at each stage. In a two-stage sampling design, a
sample of primary units is selected and then a sample of secondary units is selected
within each primary unit (Sedgwick, 2015). This technique is applicable in the study
27
considering that the students belong to different sex and year levels, who are also within
the same age group that could affect the accuracy of the data collected. According to
Kothari (2004), multistage sampling is easier to administer and can sample a large
number of units through clustering whereas this is not possible in most simple designs.
Also, the majority of the large probability samples selected from school-based to national
surveys use multistage sampling technique that includes stratification at each stage of
selection (Gilbert, 2001).
Consequently, the sample in this study was clustered by department – grade
school, junior high school, senior high school and college. Through stratified sampling,
the sample was then divided according to their grade level and sex to address the fact that
there is a wide variance in the age and ability of the respondents. A total of 125
respondents was needed in the study consisting of 45 respondents from the high school,
40 from the grade school, and another 40 from the college. This is to simulate the
students’ evacuation activity as a class. The qualified respondents were the grade 3 to 4th
year college students occupying the PM building including the 2 computer laboratories.
All respondents were physically capable and did not have any health problems.
Research Locale
The study was conducted at Holy Cross College of Calinan, a level II PAASCU
accredited Catholic school. It is located at the Davao-Bukidnon Road, Calinan, Davao
City which is the home of the Philippine Eagle and the king of the fruits, durian. The
school was founded by the PME Fathers in 1948. This is the only private sectarian school
in Calinan which has been administered by the Presentation of Mary Sisters. It envisions
28
its students to be Christ-centered, intellectually proficient, and responsive. It has four
educational buildings and it offers primary, secondary, and college education. Among the
many services, it also provides a safety plan which includes evacuation drills and safety
training.
Data Gathering Procedures
Letters of permission were sent to Holy Cross College of Calinan administration
before conducting the study. The study needed the current floor plan of the school which
was inquired from the Community Engagement and Services Office (CESO). The
distance from each source to sink, in this case, from each classroom to the evacuation
area was determined using a measuring equipment. The nodes, sources, sinks and arcs
were then identified from the floor plan by considering necessary criteria as they are
important variables in the model. It also needed data on both movement time for each arc
29
and the capacity of each node, arc, and sink through estimation methods. A speed test
was conducted to determine the movement time for each arc which included simulation
of the evacuation activity then measuring the time it would take for the class to exit the
building. The capacities of the sinks were assumed by finding the area of the location and
the estimated area of occupancy of each individual, in which according to Bappenas
(2005) is 1m2/person. In order to validate the result of the study, simulation of both the
present/non-model and proposed/model-based evacuation plan was conducted and
compared.
Data Analysis
The gathered data such as the size and measurement of building and corridors
using the floor plan, capacities of arc, nodes and sinks, and the previous evacuation time
spent using the non-model method were analyzed to get the necessary variables required
in the model. The identified arcs, nodes, sources and sinks from the floor plan of the
school were translated into a network in which the costs and the capacities of each arc,
node and sink were identified. As a prerequisite in getting the costs, the average speed
was first identified using a standard formula of speed = mean of total distance/elapsed
time. To get the cost or the movement time for each route, the standard formula: time =
distance/average speed was used. The cost was used to identify the shortest path a flow
may take to reach its sink. On the other hand, the capacities were used to assign the
sources into routes avoiding congestion which were identified through the use of the
formula: capacity = area/space occupied. The data (movement time and capacities) were
then incorporated in the Quickest Flow Model and were simulated. For validation of
results, the statistical computing program, R, was used specifically the t-test function. It
30
was used to compare before-and-after observations on the same subjects (Shier, 2004).
The time spent on evacuation both before and after the implementation of the model was
tested using the boxplot function to identify significant outliers in the data. Then, the total
egress time using the non-model-based evacuation plan was compared to the proposed
model-based plan to see if there is a significant difference between the two data
considering the normality and homoscedasticity of both data sets.

CHAPTER 3
RESULTS AND DISCUSSIONS
This chapter contains the results of the data analysis and the discussion of the
problem statements presented in the study. This presents also the interpretation of the
results and its analysis together with the references that support the response to each
research question.
Research Question #1: What is the present egress system of Holy Cross College of
Calinan?
Figure 2. Present Non-model Based Evacuation Plan of HCCC
Figure 2 shows the present egress system of Holy Cross College of Calinan not based
upon a model. This evacuation plan utilizes only one sink which is the Football Field as
destination of the evacuees and was long formulated even before the present
administration. Based on Figure 2, most of the sources were assigned to sink t1 while only
a few were assigned to the other sinks making the Football field sink congested whenever
32
occupants of the other buildings like OLSW and Rivier are present. Also, it is shown that
arc (E,G) which is the OLSW arc was not used congesting most of the flows to the stairs
of the PM building. Lastly, some of the flows were assigned incorrectly, making them more
prone to hazards and could contribute to a longer evacuation time. An example is the route
of source 10 in which it is the longest and is very dangerous because it passes below the
building itself.
Several post-disaster investigations have cited evacuation failures as a major
contributor or to the death toll during disasters (Alexander, 2002 as cited by Stepanov &
Smith, 2008). One factor is the over-utilization of a particular road link which can block
traffic through the link completely, which is in this case, the crowding of evacuees in the
stairs. This blocking will affect all upstream links which causes the significant congestion
through the evacuation network. Based on the data, the current egress system of the school
does not follow any mathematical model for optimization purposes. Also, it is only suitable
for a certain number of evacuees making it obsolete with the increased population in the
school. According to Xie (2008), many transportation and traffic related efficiency and
safety issues would be frequently encountered yet not satisfactorily addressed at the current
state of evacuation practice not following a model. In order to manage such emergencies
more effectively, decision makers may benefit from having in-place evacuation plans with
mathematical approaches for scenarios which are most likely to happen, even though real-
time design and re-evaluation of evacuation plans may be required after the disaster strikes
(Alexander, 2002 as cited by Stepanov & Smith, 2008).

33
Research Question #2: What is the average time spent during evacuation using the non-
Table 1. Time spent using the non-model evacuation plan with observations
Drill no. Date Observations Elapsed Time
No. 1 February 27, The guidance sink was not used. 4 minutes and 1
2017 There was a heavy congestion in both second
(replace (Afternoon stairs A and B. The OLSW arc was
ment of Break) not used. Most students were just
the walking.
outlier)
No. 2 July 6, 2017 The guidance sink was already used. 4 minutes and 5
(Morning The students in the football sink were seconds
Break) blocking the entry place of
emergency vehicle. The OLSW arc
was not used resulting in a heavy
clogging in both stairs A and B.
No. 3 July 11, 2017 The students in both sinks were not 4 minutes
(Morning arranged properly. Students from
Break) source 10 followed a dangerous route
which is across the building. Still the
OLSW arc was not used resulting in
a heavy clogging in both stairs A and
B.
No. 4 July 14, 2017 Students did not walk faster than 4 minutes and 6
(Afternoon normal, instead some just walked seconds
Break) towards the evacuation areas.
No. 5 July 20, 2017 The students in both sinks were not 3 minutes and 54
(Morning arranged properly. Students from seconds
Break) source 10 followed a dangerous route
OLSW arc was not used resulting in
a heavy clogging in both stairs A and
B. Some students were not in their
rooms. Others were running instead
of walking faster than normal.
34
Table 1 provides the observations and the time spent to evacuate the building using
the non-model based evacuation plan with an average of 4 minutes and 1 second. During
the analysis of the data set, it was identified that the fourth drill conducted on July 18, 2017
is a significant outlier using the boxplot function which could affect the result of the study.
To address such problem, the researchers used the evacuation time conducted last school
year on February 27, 2017 since conducting another drill with the non-model based plan is
not anymore possible. It is shown that the fourth drill conducted on July 15, 2017 had the
longest time which took the evacuees 4 minutes and 6 seconds to evacuate while the fifth
drill conducted on July 20, 2017 took the shortest time with only 3 minutes and 54 seconds.
Different factors such as the location of the evacuees, congestion along the stairs and the
drill behavior affected the overall evacuation time. The problems observed are repetitive;
however, they are still not addressed in the non-model based evacuation plan.
Evacuation by nature is a transportation activity involving traffic. The congestion
generated along the stairs may be a result of a surge of traffic demand exceeding the
capacity of the arc (Kalafatas & Peeta, 2009). The problem involving the exit path of
source 10 and the congestion of evacuees needs to be dealt with properly as an evacuation
plan must be enacted with the aim of reducing the life-threatening risk of the susceptible
population to minimum level (Xie, 2008). Also, the evacuee behavior which is a complex
phenomenon should be understood since it causes most of the clogs during evacuations
(Stepanov & Smith, 2008). The undesirable behavior of the evacuees like not walking
faster than normal, shouting and laughing may contribute to the long evacuation time. That
is why by using an evacuation model, though it does not cover the behavior of evacuees
under emergency conditions, it will update the evacuation routes in time periods which will
35
minimize the overall evacuation time (Khadka, 2016) despite the different behaviors that
the evacuees are exhibiting.
Research Question #3: Is the QFM applicable for planning an evacuation strategy in
Holy Cross College of Calinan?
Table 2. The applicability of the mathematical model based on the demands of the
school’s present egress system
Problems in the present egress system of Quickest flow problem approaches

Holy Cross College of Calinan
Does not follow a model-based plan Network modelling with mathematical

approaches
Arrangement and distribution of students to Sink Capacity and Sink Assignment

different evacuation areas.
Longer path taken by some students Shortest Path Algorithm
Long time completion of evacuation activity Objective function: T

min ∑∑txid (t)
t=0 iϵD
Congestion of students in the stairs Arc Capacity
Misutilization of available arcs and Source Distribution and Arc Congestion

pathways
Table 2 shows the different problems present in the current evacuation plan of the
school under study along with the different approaches that the mathematical model -
Quickest Flow Problem, offers. Based on Table 1, it was identified that there were existing
problems observed during the simulation of evacuation drills. These problems include no
proper arrangement and distribution of students in the evacuation areas and flawed route
assignment in which some flows take longer paths which lead to a longer evacuation time.
36
Also, the total time exceeds the ideal time set by the Bureau of Fire Protection (BFP) which
is only 3 minutes caused by the congestion of students in certain areas specifically the stairs
and the improper use of pathways leaving some arcs empty during the evacuation activity.
In this connection, the QFM offers different approaches to address those problems at the
same time that provides a basis for the formulation of evacuation plan using mathematical
processes. The following approaches were offered: first, it considers the optimal
arrangement and distribution of students in the evacuation areas considering their
capacities to promote an organized evacuation activity. Second, it uses Shortest Path
Algorithm specifically Djkistra’s which aims to route students to paths with shorter
distances. Through the model’s objective function:
T
min ∑∑txid (t)
t=0 iϵD
It aims to lessen the time horizon of an evacuation activity. The model also considers the
capacities of arcs and sinks and the distribution of students at each time interval. This is to
avoid traffic congestions during the evacuation activity. Lastly, the QFM routes the
evacuees to paths with shorter traversal times and utilizes available arcs to lessen the total
evacuation time.
Basically, evacuation is defined as the removal of lives and/or property from the
disaster zone to the safety zone as quickly as possible (Khadka, 2016) and minimizing the
evacuation time by which all evacuees completed the evacuation is the criterion for the
evacuation planning (Katoh, et al., 2011). In order to minimize the time of evacuation,
different strategies should be employed like the identification of the shortest path,
37
minimizing traffic congestion, identifying path cost and capacities and the formulation of
proper route distribution. Though the development of such strategies is a difficult task, it
may be greatly aided through involving mathematical modeling with optimization insights
(Zheng, et al., 2010) like the Quickest Flow Problem (QFP). According to Xie (2008), a
well-defined and manageable plan based on network models is one of the prerequisites for
successful implementation of a large-scale evacuation. The QFP aims to send a given flow
amount from the source to the sink in the shortest possible time. The resulting minimal
time T therefore gives a lower bound on the minimal evacuation time (Hamacher, et al.,
2011). Since HCCC does not use a network model in its current evacuation plan, providing
one would definitely improve the egress system and give confidence to the evacuees to
follow a plan knowing that it is based on comprehensive research.

38
Research Question #4: What procedures should be employed to improve egress
efficiency?
Figure 3.1. Shortest Path Algorithm of Sources 1 and 2
Figure 3.1 shows the set of paths sources 1 and 2 that may take towards the different
evacuation areas. Through Shortest Path Algorithm, it determines the shortest distance a
source may travel among the identified set of paths to reach the nearest evacuation area.
The costs or the time it takes to reach a node through an arc are shown in time units in
which one time unit is equal to 10 seconds. Based on Figure 3.1, the shortest path sources
1 and 2 may take is from S to arc (S,A), to arc (A,B), to arc (B,C), then all the way to arc
(C,t2) to reach the evacuation area at the shortest possible time of 12 time units or
approximately 2 minutes compared to other paths which may lead to longer evacuation
times.
39
Figure 3.2. Shortest Path Algorithm of Sources 3, 4 and 5
Figure 3.2 shows the set of paths sources 3, 4 and 5 that may take towards the
different evacuation areas. Based on the figure, the shortest path they may take is from S
to arc (S,B), to arc (B,C) then all the way to arc (C,t2) to reach the evacuation area at the
shortest possible time of 10 time units or approximately 1 minute and 40 seconds.
Figure 3.3. Shortest Path Algorithm of Sources 6, 7, 8, 9 and 10
Shown in Figure 3.3 is the set of paths sources 6, 7, 8, 9 and 10 that may take
towards the different evacuation areas. Based on the figure, the shortest path they may take
is from S to arc (S,E) to arc (E,G) then all the way to arc (G,t1) to reach the evacuation
area at the shortest possible time of 14 time units or approximately 2 minutes and 10
seconds.
40
In Figure 3.4, it can be noted the set of path sources 11 and 12 that may take towards
the different evacuation areas. In additional, the shortest path they may take is from S to
arc (S,C) then all the way to arc (C,t2) to reach the evacuation area at the shortest possible
time of 6 time units or approximately 1 minute.
Figure 3.5. Shortest Path Algorithm of Source 13
Additionally, Figure 3.5 shows the set of paths source 13 that may take towards the
different evacuation areas. The figure clearly outlines the shortest path source 13 may take
is from S to arc (S,C) then all the way to arc (C,t2) to reach the evacuation area at the
shortest possible time of 7 time units or approximately 1 minute and 10 seconds.

41
Figure 3.6. Shortest Path Algorithm of Sources 14, 15 and 16
The set of paths sources 14, 15 and 16 may take towards the different evacuation
areas as shown in Figure 3.6. Clearly, the shortest path they may take is either from S to
arc (S,C) then all the way to arc (C,t2) or from S to arc (S,G) then all the way to arc (G,t1)
to reach the evacuation area at the shortest possible time both of 10 time units or
approximately 1 minute and 40 seconds.
Figure 3.7 shows the set of paths sources 17 and 18 that may take towards the
different evacuation areas. Here, the shortest path they may take is from S to arc (S,G) then
all the way to arc (G,t1) to reach the evacuation area at the shortest possible time of 8 time
units or approximately 1 minute and 20 seconds compared to other paths which may lead
to longer evacuation times.

42
Before formulating a new strategic evacuation that improves egress efficiency,
observations of the current evacuation system happened first. One problem that was
identified within those observations included the lack of proper route assignment affecting
the total time of evacuation. Evacuees were not properly assigned to routes where they can
safely pass during an evacuation. This affects the flow of the students while going out from
the building. The problem contributes to the time of evacuation because congestions
usually happen during the evacuation process (Khadka, 2016). In order to answer the given
problem, finding the shortest path that the maximum flow could attain is an advantage
(Stepanov & Smith, 2008). Using the Shortest Path Algorithm that was shown in the figures
above, the shortest paths that cover the shortest time to evacuate were identified (Kerbache
& Smith, 2000 as cited by Stepanov et al., 2008). From different sets of sources, each
source was shown to proceed to assigned sinks where it covers the shortest time to
evacuate. The given time units serve as the time of evacuation following the path going to
different sink in which one-time unit is equal to ten seconds. Therefore, the bigger the time
unit, the more time it will take to travel from a source to a sink.
43
Table 3. Costs of different arcs in time units
Arc Cost
(A, B) 3 time units
(B, C) 3 time units
(C, t2) 4 time units
(C, I) 7 time units
(I, t1) 5 time units
(E, F) 3 time units
(D, E) 3 time units
(E, G) 4 time units
(F, t3) 6 time units
(F, H) 6 time units
(H, t1) 6 time units
(G, t1) 5 time units
Table 3 shows the costs of the different arcs represented in time units with one unit
equal to ten seconds. The arcs with the least amount of cost are arcs (A, B), (B, C), (E, F),
and (D, E). Each of these four arcs requires three (3) time units or approximately 30 seconds
for a single flow to pass through. On the other hand, the arcs with the greatest amount of
cost are arcs (C, t3) and (C, I) with each requiring 7 time units or approximately 70 seconds
for a flow to traverse. Therefore, travelling from node C to t3 and C to I doubles the time
units than travelling from node A to B. These are transit times or the time it takes for a
single flow to travel from the starting point to the end point of an arc. It is determined by
the arc’s length and the average speed taken from the speed test conducted in the research.
44
Through these data, the researchers were able to route sources effectively to lessen the
overall time.
According to Katoh, et al. (2011), minimizing the evacuation time by which all
evacuees completed the evacuation is the criterion for the evacuation planning. This can
be achieved by identifying the cost of different arcs to be used during evacuation.
Therefore, an efficient evacuation strategy will present routes with the least amount of time
it takes to travel (Pyakurel & Dhamala, 2014). With that, the smaller the arc cost, the faster
it allows a flow to traverse.

45
Table 4. Capacity of different arcs relative to flow size
Arc Capacity
(A, B) 2 flows
(B, C) 2 flows
(C, t2) 3 flows
(C, I) 5 flows
(I, t1) 4 flows
(E, F) 2 flows
(D, E) 2 flows
(E, G) 3 flows
(F, t3) 5 flows
(F, H) 5 flows
(H, t1) 5 flows
(G, t1) 3 flows
(C, t3) 5 flows
Table 4 shows the capacities of the different arcs relative to the flow size of each
source. It was shown in flows to capture the crucial element that routing occurs over time
(Fleisher & Skutella, 2007). For example, two (2) flows can travel from node A to B, and
from node B to C, at the same time. However, arcs will have different capacities depending
on its area size. Some arcs may have greater capacity than the others. Meaning, in
considering the route assignment of the sources, it is important to consider the arc
capacities to avoid congestion.
An arc having greater capacity is as valuable as having lesser travel time. During
an earthquake disaster, it is expected that mass evacuation is performed immediately in

46
order to avoid dangers resulting in the disaster. According to Cruz, Talebi and Smith (2005)
as cited by Stepanov and Smith (2008), the improvement of the performance depends on
the efficient management of the demand-supply problem of the evacuation network. In this
case, arc capacity is very much important to consider. In fact, Zimmerman (2010) said that
the efficient and expeditious flow of evacuation traffic is the most crucial element in a
successful evacuation. However, it is also the most challenging mission as the over-
utilization of a particular road link or arc can block traffic through the link completely if
exceeding the arc capacity. This blocking will affect all upstream links which causes the
significant congestion throughout the evacuation network (Xie, 2008).
Table 5. Capacity of different sinks based on the average occupancy
Sink Area Safe Zone Capacity
Football Field (t1) 1,158.4485 m2 891 m2 891 persons
Front of Guidance (t2) 524.8615 m2 364 m2 364 persons
Pieta (t3) 317.4233 m2 143 m2 143 persons
Table 5 shows the capacities of the different sinks based on the average occupancy
of every individual. In the study, three (3) possible areas that can be used as evacuation
sinks for the PM building occupants were identified, namely: the Football field, the front
of Guidance Offices and Pieta. By measuring the dimensions and partitioning, the different
areas were identified as shown in the table 5. With the hazards present around these areas
like trees, pathways, and walls, an amount of space was allotted away from these hazards
thus lessening the area which is called the safe zone. From these safe zones, each capacity
47
was identified using the formula Capacity = area/ average occupancy. Based on the results,
a total of 2,218 evacuees could fit in the identified evacuation areas.
According to Saadatseresht, et al. (2009), a central challenge in developing an
evacuation plan is determining the distribution of evacuees to where they should go, that
is, the evacuation area or sinks. All sinks must be away from dangerous and adverse
hazards like water, low shrubs, areas covered by trees, non-refuge buildings and slopes
greater than 15 degrees (Wang, et al., 2013). That is why in considering a safe zone,
allotting a space from each hazard is highly required. By considering those hazards, the
safety of the evacuees will be ensured which is the main purpose of having an evacuation
plan: to promote safety and to save lives. Also, he added that the evacuation areas should
have sufficient reserved space to accommodate more people during evacuation, which is
according to Bappenas (2005), it is needed to have a designated space of 1m2 per person to
avoid congestion in the sinks. That is why, in identifying the capacity of each sink, each
safe zone is divided by 1m2 which serves as the average occupancy of each individual.
48
Table 6. Distribution of flows at each time interval
ARC TIME INTERVAL
A-B 0-29.09s 29.09 – 58.18 s 58.18 – 87.2 s
Arc TIME INTERVAL
B-C 0 – 29.09 s 29.09 – 58.18 s 58.18 – 87.2 s
2 3
Arc TIME INTERVAL
E-F 0 – 28 s 28 – 56 s 56 – 84 s
ARC TIME INTERVAL
E-G 0 – 45.74 s 45.74 – 91.48 s 91.48 s – 137.22 s
2 1
ARC TIME INTERVAL
G-T1 0 – 49.6 s 49.6 – 99.2 s 99.2 – 148.8 s
3 3
ARC TIME INTERVAL
C-T2 0 – 42.48 s 42.48 – 84.96 s 84.96 – 127.44 s
2 4
ARC TIME INTERVAL
C-I 0 – 69 s 69 – 138 s 138 – 207 s
ARC TIME INTERVAL
F-H 0 – 66.52 s 66.52 – 133.04 s 133.04 – 199.56 s
ARC TIME INTERVAL
I-T1 0 - 52.32 s 52.32 – 104.64 s 104.64 – 156.96 s
ARC TIME INTERVAL
H-T1 0 - 66.52 s 66.52 – 133.04 s 133.04 – 199.56 s
4
49
Table 6 shows the distribution of flows in arcs at different time intervals. Some arcs
exceeded the capacity at certain time intervals like the arc (B,C) at 58.18 seconds to 87.2
seconds which contains 3 flows exceeding the capacity of the arc which is 2. It is the same
with the arc (C,t2) at 1.25 minutes to 2.07 minutes which contains 4 flows exceeding the
arc capacity which is only 3 flows. These are due to the surge of traffic demand in the right
wing of the building with only one possible exit.
Kalafatas and Peeta (2009) also noticed the same major setback of evacuation
problem which might exceed the capacity of an existing system and results in congestion.
Moreover, Zimmerman (2010) also mentioned that these heavy flows can be results from
a rapid rise of demand. Queues and delays are expected under such scenarios, as the
demand volume or the flow exceeds the capacity of the arc. The capacity of the arc limits
the rate of flow at each point in time. Though this problem cannot be addressed, it is
considered the better approach than re-routing the sources which may lead to a more
congested evacuation plan.

50
Figure 4. Model-based Evacuation Plan
Figure 4 shows the formulated evacuation plan based on QFM. It considers
different factors like the distance, capacities of arcs and sinks, arc costs and the distribution
of flows at each time interval. Moreover, it lessens the possibility of congestions, routes
sources to paths with less distances, lessens the overall evacuation time, utilizes all
available paths to avoid grid-lock of the model, and organizes the flow of students from
the sources up to their specific arrangements in the evacuation areas. Compared to the non-
model based evacuation plan, the different sources are routed and distributed properly
utilizing all available paths.
Evacuation procedures in buildings have long been studied in many aspects from
architecture, engineering, computer modeling, etc (Chizari, et al. 2013). Despite the
numerous efforts, evacuating a building keeps getting complicated as time goes by.
Without proper egress system, evacuation may lead to problems that will result in more
casualties than expected when a calamity such as an earthquake occurs. Furthermore, in

51
order to manage such emergencies effectively, an efficient evacuation plan must be
addressed (Alexander, 2002 as cited by Stepanov & Smith, 2008) using a simulation model,
or in other words, in these cases, an integrated optimization-simulation approach is highly
desirable (Xie, 2008). According to Xie (2008) and Khadka (2016), many factors should
be considered when formulating an evacuation plan. These factors should be addressed
properly since they affect the overall efficiency of the evacuation plan which could be
assessed by evaluating the number of evacuees having left the disaster area over time.
Figure 5.1. Assignment of Evacuees in the Football field Sink
Figure 5.1 shows the assignment of the different sources in the Football field sink
or sink t1. A total of 12 sources which are approximately 540 students coming from the
52
PM building were assigned to their respective areas in the field in which the sources that
reach the sink first are assigned to the far side or corner with a space allotment of 45 m2
each. Though the study is only concerned with the occupants of the PM building, still the
occupants of the other buildings are also considered since they could affect and contribute
to the congestion in the different sinks during real-time evacuation activity. The occupants
of the College and OLSW buildings are assigned to the right side of the field with an
approximate number of 280 evacuees. Considering the capacity of the sink, the assigned
evacuee population which is approximately 820 does not exceed the capacity which is 891
persons, thus lessening the possibility of congestion.
Figure 5.2. Assignment of Evacuees in Front of Guidance offices
In sink t2 fronting the guidance offices, a total of 6 sources which are about 270
evacuees from the PM building were assigned. Similar with sink t1, the sources that arrive
the sink first are assigned to the far side of the area. Considering the capacity of the sink,
53
the assigned evacuee population which is approximately 270 does not exceed the capacity
of 364 persons.
Figure 5.3. Assignment of Evacuees in Pieta Sink
In sink t3 at the Pieta area, only the occupants of the different offices and HE
located at the ground floor of the PM building were assigned. This is because the path itself
towards the sink is a major hazard for those flows coming from the upper floors.
Considering the capacity of the sink, the assigned evacuee population which is only
approximately 20 does not exceed the capacity which is 143 persons.
To promote an organized evacuation, the evacuation process must be specified and
fixed up to the end, that is, the assignment of sources in the different sinks. According to
54
Karnema, et al. (2014), for safety purposes, evacuation zones should be established before
a disaster happens. This includes the identification of boundaries for the safe zones and the
elimination of the possible hazards in the areas.. On the other hand, evacuees should also
be notified of their location in advance to reduce evacuation related risks and costs. By
providing a specific place for the evacuees to stay during disasters, it may result in a more
organized evacuation activity and has the potential to improve evacuation efficiency
greatly (Yuan, et al., 2006).

55
Research Question #5: What is the average time spent during evacuation using the
Table 7. Time spent using the model-based evacuation plan with observations
No. 1 September 6, 2017 The students were able to follow 3 minutes and
(Afternoon Break) the new evacuation plan but many 22 seconds
were just walking towards the
sink. There was less congestion in
the football sink.
No. 2 September 7, 2017 Stair B became less congested and 3 minutes and
(Afternoon Break) occupants in the left wing arrived 25 seconds
at the sink with less time. Most
students passing the OLSW path
were just walking. Many Grade 7
students were playing and
laughing. There was a congestion
in OLSW path due to the DAT.
No. 3 September 9, 2017 Congestion due to the DAT was 3 minutes and
(Morning Break) addressed. Many students were 20 seconds
still walking and playing. Students
in the sinks were arranged and
organized.
No. 4 September 19, 2017 Still, a smooth flow can be 3 minutes and
(Morning Break) observed in Stair B. Occupants of 28 seconds
the AVR criss-crossed with other
evacuees causing congestion.
Many students were walking.
No. 5 September 20, 2017 Only a few students can be 3 minutes and
(Morning Break) observed walking. The presence of 15 seconds
DAT hastened the evacuation
activity. There was less congestion
in both stairs A and B.
Table 7 provides the observations and the time spent to evacuate the building using
the model-based evacuation plan with an average of 3 minutes and 22 seconds. The third
drill conducted on September 19, 2017 took the longest time with 3 minutes and 28 seconds
56
to evacuate while the fourth drill conducted on September 20, 2017 took the shortest with
only 3 minutes and 15 seconds. Different factors such as proper route assignment, evacuee
arrangement in sinks, lessened possibility of congestion and presence of the administrators
improved the overall evacuation time. Despite the negative behavior of the students during
the drills, employing an evacuation plan based on mathematical model deemed to have
evident positive results which reduced the overall previous evacuation time of up to 45
seconds.
According to Xie (2008), evacuation planning may deal with both short-term and
long-term issues like increase in evacuee population and availability of path and sinks. For
a long-term evacuation plan, it is generally proposed for a potential emergency area in
which some natural disasters may have frequently occurred in history and are expected to
occur again in a foreseeable future. By addressing both the short-term and long-term issues
in the previous evacuation plan through the mathematical model, the over-all time may be
reduced. Also, evacuee behavior plays a big role in the evacuation activity. With the new
evacuation model, still undesirable behaviors like not walking faster than normal, laughing
and shouting were present. But according to Shen (2005) as cited by Chizari, et al. (2013),
the social and familiarity factors of evacuees can help address these behaviors in a real-
time evacuation activity in which the assurance due to having an efficient evacuation plan
will make the evacuees follow their assigned routes to proceed safely to the evacuation
areas.
57
Research Question #6: Is there a significant difference in the time spent between the
non-model and model-based evacuation plan?
Table 8. Determining the difference using t-test
Test Variable 1: Non-model- Variable 2: Model-based

based evacuation plan evacuation plan
mean 241.2 202
t stat (computed values) 12.7585
t critical 2.3060
p-value 1.34208e-06
It is shown in table 8 the statistical results of the two data sets which are the
evacuation time before and after the implementation of the Quickest Flow Model. It can be
noted that the t-stat (computed) of 12. 7585 is greater than the t-critical having a value of
2.3060. This implies that there is a significant difference in the evacuation time when a
model was used as compared with the one without the model. This is supported with the p-
value of 1.34208 -06 which is lower than the alpha value of 0.05. This means that using
the new evacuation plan, the overall time spent of evacuation was reduced.
According to Hooks and Patterson (2004) as cited by Baumann and Skutella (2009),
the importance and wide applicability of network flow problems has long been recognized.
Dynamic network flow problems, in particular, have been used to model numerous real
world phenomena arising in applications within almost all industries such as production-
distribution systems, fleet management, evacuation and communications. The QFM in
particular, aims to send a maximum amount of flow from a source to a sink within a given
time bound. This routes the evacuees to safety with the least possible time (Pyakurel &
58
Dhamala, 2014). This was supported by Hamacher, Leiner, and Ruzika, (2011) when they
mentioned that an optimal dynamic network flow can be computed efficiently and yields a
best achievable evacuation plan, provided all evacuees follow the routes corresponding to
this flow.
CHAPTER 4
CONCLUSION AND RECOMMENDATIONS
This chapter contains the conclusions and recommendations drawn from the data
analyzed in the previous chapter.
Conclusion
In this paper, different approaches both mathematical and strategic were
considered and employed to provide an evacuation plan that would lessen the overall
evacuation time of the school. Through Djkistra’s Algorithm, the shortest paths were
determined by comparing the distances of all the possible routes from each source to the
nearest evacuation area. The capacities of arcs and sinks were also identified by
computations through the use of standard formulas and estimation methods. This would
help route flows efficiently using available and uncongested pathways which would
lessen the possibility of congestions both in the arcs and in the sinks during real-time
evacuation activity. Another factor considered is the supply and demand factors,
evacuees were assigned and arranged in specific locations in the evacuation areas which
helped in organizing the crowd of evacuees. Lastly, the most important variable which is
time expressed in time units was also captured in the model by foreseeing the distribution
of evacuees in the different arcs at each time interval and computing the evacuation time
through the model’s objective formula given the different costs and the speed of the
evacuees.
60
Recommendations
Based on the scope and delimitations of the study, several recommendations were
made and should be considered by future researchers.
First, since the newly imposed routing strategy significantly decreased the total
time of evacuation, it is recommended that the proposed strategy should be followed as
the new school’s evacuation plan. Second, given the constant changes in the population
in the school, it would be better to incorporate both previous and forecasted enrollment
data as the assume number of evacuees. Third, it is recommended to consider different
behaviors of the evacuees under emergency conditions like panic, herding, congestion,
counter flows and ignorance of guidance which could greatly affect the evacuation
activity despite the presence of an evacuation plan. Fourth, since the study only provides
evacuation routing strategy for the PM building, it is advantageous as well for the future
related studies to provide evacuation routes for the Rivier, Our Lady Seat of Wisdom and
Elementary buildings. Fifth, it is also recommended to increase the number of
simulations or evacuation drills to address the possible variance in the data when
determining whether the time of evacuation decreased significantly after providing the
model-based routing strategy. Sixth, future researchers may also consider the presence of
the secondary disasters that may occur like fire which could affect the flow of evacuees if
certain arcs and nodes will be affected. This is the same with unavailability of certain
routes in the building due to various factors and the possibility of blockage in the exits.
Lastly, other researchers may also venture on developing the disaster response strategy as
it is also part of the emergency preparedness plan of the school.

61
References
Ahuja, R., Magnanti, T., & Orlin, J. (1993). Network flows: theory, algorithms, and
applications. Retrieved from https://pws.yazd.ac.ir
Ali, M. M., & Moon, K. S. (2007). Structural developments in tall buildings: Current trends
and future prospects. Architectural Science Review, 50(3), 205-223.
doi:10.3763/asre.2007.5027
Amirgaliyeva, Z. (2012). Evaluation of the importance of mathematical modelling in

finance through a discussion of its uses and misuses. Opticon1826(12), 12-13.
BAPPENAS (2005). Indonesia: preliminary damage and loss assessment (W. Bank,
Trans.). Indonesia. Retrieved from https://www.bappenas.go.id/
Baumann, N., & Skutella, M. (2009). Solving evacuation problems efficiently: earliest
arrival flows with multiple sources. Mathematics of Operations Research, 34(2).
Burtles, J., & Noakes-Fry, K. (2016). The challenge of emergency evacuation from high-
rise buildings. Rothstein Associates, Incorporated, 2014.
Castle, C. J. E. (2006). Developing a prototype agent-based pedestrian evacuation model

to explore the evacuation of King's Cross St. Pancras underground station.
Caunhye, A. M., Nie, X., & Pokharel, S. (2011). Optimization models in emergency
logistics: A Literature review.
Chang, L. (2010). Transportation system modeling and applications in earthquake

engineering. Univeresity of Illinois at Urbana-Champaign.
Chen, J.-C. (2003). Dijkstra’s shortest path algorithm. Journal of Formalized Mathematics,
15(9).
Chizari, H., Malekinezhad, F., Embi, M. R., Yatim, Y. M., Razak, S. A., Ahmad, M. H. b.
H., & Bakhtiara, M. (2013). Agent-based approach for modeling evacuee uncertainty
behavior using game theory model. Life Science Journal, 10(3), 1350-1355.
Dawson, R. (2011). How significant is a boxplot outlier. Journal of Statistics Education,

19(2), 1-12.
Estrada, E. (2013). Graph and network theory in physics. Retireved from https://arxiv.org
Fleischer, L., & Skutella, M. (2007). Quickest flows over time. SIAM Journal on
Computing, 6(36), 1600-1630.
62
Gilbert, N. (2001). Research, theory and method. Researching social life, 2. Retrieved from
https://epubs.surrey.ac.uk/1586/1/fulltext.pdf
Group, O. B. (2012). The report: the Philippines 2012 Country Profile (pp. 11).
Hall, J. R. (2011). High-rise building fires. National fire protection association.
Hamacher, H. W., Leiner, K., & Ruzika, S. (2011). Quickest cluster flow problems.
Pedestrian and Evacuation Dynamics.
Hammer, Ø., Harper, D., & Ryan, P. (2008). PAST-palaeontological statistics, ver. 1.89.
Paleontological Museum, University of Oslo, Noruega.(También disponible en línea:
http://folk. uio. no/ohammer/past/index. html).
Hosseini, M., & Izadkhah, Y. (2006). Earthquake disaster risk management in schools.
Disaster Prevention and Management: An International Journal, 15(4), 649-661.
Ibrahim, E. (2007). High-Rise Buildings–Needs & Impacts. Paper presented at the CIB
World Building Congress.
Kalafatas, G., & Peeta, S. (2009). Planning for Evacuation: insights from an efficient
network design model. Journal of Infrastructure Systems, 15(1), 21-30.
Karnema, A., Mohadeseh, H., Samira, H., & Asghari, Z. (2014). A optimal model for
emergency evacuations in an urban areas (case: Kerman).
Katoh, N., Takizawa, A., & Inoue, M. (2012). An emergency evacuation planning model
based on the universally quickest flow.
Khadka, S. R. (2016). Optimal Traffic Planning for Efficient Evacuation. Journal of

Advanced College of Engineering and Management, 1, 119-126.
Kothari, C. R. (2004). Research methodology methods and techniques (Second Revised

ed.).
Larget, B. (2014). Chapter 2 R ggplot2 Examples Lock 5.
Li, W., Zhu, J., Li, H., Wu, Q., & Zhang, L. (2015). A game thoery based on monte carlo
analysis for optimizing evacuation routing in complex scenes. Mathematical Problems
in Engineering. doi:10.1155/2015/292093
Lindell, M. K., & Prater, C. S. (2007). Critical behavioral assumptions in evacuation time
estimate analysis for private vehicles: Examples from hurricane research and planning.
Journal of urban planning and development, 133(1), 18-29.
63
Lo, S., Huang, H., Wang, P., & Yuen, K. (2006). A game theory based exit selection model
for evacuation. Fire safety journal, 41(5), 364-369.
Lu, Q., Huang, Y., & Shekhar, S. (2003). Evacuation planning: a capacity constrained
routing approach. Intelligence and Security Informatics, 959-959.
Lu, X., Lu, X.Z., & Zhang, W.K. (2011). Collapse simulation of a super high-rise building
subjected to extremely strong earthquakes. Sci.China Tech Sci., 10.
doi:10.1007/s1431-011-4548-0
Ma, J., Song, W., Tian, W., Lo, S. M., & Liao, G. (2012). Experimental study on an ultra
high-rise building evacuation in China. Safety Science, 50(8), 1665-1674.
Moore, D. S. (2007). The basic practice of statistics (Vol. 2). New York: WH Freeman.
Nazarov, E. (2011). Emergency response management in Japan. Azerbaijan: Crisis

Management Center, Ministry of Emergency Situations of the Republic of Azerbaijan.
Ng, C., & Chow, W. K. (2006). A brief review on the time line concept in evacuation.
International journal on architectural science, 7(1), 1-13.
Nilsson, D., & Jonsson, A. (2011). Design on evacuation systems for elevator evacuation
in high rise buildings. Journal of Disaster Research, 6(6), 600-609.
Peeta, S., Sharma, S., & Hsu, Y.-T. (2011). Dynamic real-time routing for evacuation
response planning and execution.
Pel, A., Bliemer, M., & Hoogendoorn, S. (2011). A review on travel behavior modelling
in dynamic traffic simulation models for evacuations. Transportation Journal.
doi:10.1007/s11116-011-9320-6
Pelechano, N., & Badler, N. (2006). Modeling crowd and trained leader behavior during
building evacuation (Vol. 26, pp. 80-86).
Pyakurel, U., & Dhamala, T. N. (2014). Earliest arrival contraflow model for evacuation
planning. Neural, Parallel, and Scientific Computations, 22, 287-294.
Qui, K.-F., & Jin, W.-L. (2008). Studies of emergency evacuation strategies based on
kinematic wave models of network vehicular traffic. Intelligent Transportation
Systems.
Revita, J. (2017). Dabawenyos evacuate due to 7.2 earthquake. SunStar Davao.
Ronchi, E., & Nilsson, D. (2013). Fire evacuation in high-rise buildings: a review of human
behavior and modelling research.
64
Saadatseresht, M., Mansourian, A., & Taleai, M. (2009). Evacuation planning using
multiobjective evolutionary optimization approach. European Journal of Operational
Research, 198(1).
Sagun, A., Anumba, C., & Bouchlaghem, D. (2013). Designing buildings to cope with
emergencies: findings from case studies on exit preferences. Buildings, 3(2), 442-461.
Schmitt, J. F. (2006). A systemic concept for operational design. US MC Warfighting

Laboratory. http://www. mcwl. usmc. mil/file_download. cfm.
Sedgwick, P. (2015). Multistage sampling. doi:10.1136/bmj.h4155
Sena, L., & Woldemichael, K. (2006). Disaster prevention and preparedness. Ethopia
Public Health Training Initiative.
Shen, Z.-J. M., Pannala, J., Rai, R., & Tsoi, T. S. (2008). Modeling transportation networks
during disruptions and emergency evacuations. University of California
Transportation Center.
Shier, R. (2004). Statistics: 1.1 Paired t-tests. United Kindom: Mathematics Learning
Support Centre.
Sieh, K. (2007). The Sunda megathrust: past, present and future. Journal of Earthquake
and Tsunami, 1(1), 1-19.
Sivakumar, S., & Chandrasekar, C. (2014). Modified Dijkstra’s shortest path algorithm.
International Journal of Innovative Research in Computer and Communication
Engineering, 2(11), 6450-6456.
Sons, J. W. (2005). Quasi-experimental design Encyclopedia of biostatistics.
Spearpoint, & Maclennan. (2012). The effect of an ageing and less fit population on the
ability of people to egress buidlings. Safety Science, 50.
Steen, M. v. (2010). Graph theory and complex networks. An Introduction, 144.
Stepanov, A., & Smith, J. M. (2008). Multi-objective evacuation routing in transportation

networks. European Journal of Operational Research, 198(2), 435-446.
doi:10.1016/j.ejor.2008.08.025
UNESCO (2012). Plan of action for safe school and educational buildings in Khyber
Paktunkwa. Pakistan.
Wang, Z., Wang, X., Liu, Y., & Chu, J. (2013). Study on methods of urgent refuge planning
based on GIS. Applied Mechanics and Materials.
doi:10.4028/www.scientific.net/AMM.353-356.2251
65
Xie, C. (2008). Evacuation network optimization models, solution, methods and

applications (Doctoral Dissertation), Cornell Univeresity.
Xie, C., Lin, D.-Y., & Waller, S. T. (2010). A dynamic evacuation network optimization
problem with lane reversal and crossing elimination strategies. Transportation
Research part E: logistics and transportation review, 46(3), 295-316.
Yang, W. B., Han, S. W., Zhang, J. J., & Song, W. (2004). Planning construction of
earthquake emergency shelters and urban disaster reduction. Journal of Natural
Disasters, 13(1), 126-131.
Yuan, F., Han, L., Chin, S.-M., & Hwang, H. (2006). Proposed framework for
simultaneous optimization of evacuation traffic destination and route assignment.
Transportation Research Record: Journal of the Transportation Research
Board(1964), 50-58.
Yuan, J. S., Reed, A., Chen, F., & Stewart, C. N. (2006). Statistical analysis of real-time
PCR data. BMC bioinformatics, 7(1), 85.
Zheng, H., Chiu, Y.-C., Mirchandani, P. B., & Hickman, M. (2010). Modeling of
evacuation and background traffic for optimal zone-based vehicle evacuation strategy.
Transportation Research Record: Journal of the Transportation Research
Board(2196), 65-74. doi:10.3141/2196/07
Zimmerman, C. (2010). Using Highways for No-Notice Evacuations: Routes to Effective

Evacuation Planning Primer Series: DIANE Publishing.
Appendices
Appendix 1a. Time spent using the non-model evacuation plan with observations

No. 1 February 27, The guidance sink was not used. 4 minutes and 1
2017 (Recess There was heavy congestion in both second
(replacem Afternoon) stairs A and B. The OLSW arc was
ent of the not used. Most students are walking.
outlier)
No. 2 July 6, 2017 The guidance sink was already used. 4 minutes and 5
(Recess The students in the football sink were seconds
morning) blocking the entryplace of emergency
vehicle. The OLSW arc was not used
resulting in a heavy clogging in both
stairs A and B.
No. 3 July 11, 2017 The students in both sinks were not 4 minutes
(Recess arranged properly. Students from
Morning) source 10 followed a dangerous route
OLSW arc is not used resulting to
heavy clogging in both stairs A and B.
No. 4 July 14, 2017 Students did not walk faster than 4 minutes and 6
(Recess normal, instead some just walked seconds
Afternoon) towards the evacuation areas.
No. 5 July 20, 2017 The students in both sinks are not 3 minutes and 54
(Recess arranged properly. Students from seconds
Morning) source 10 follow a dangerous route
OLSW arc is not used resulting to
heavy clogging in both stairs A and B.
Some students were not in their
rooms. Others were running instead
of walking faster than normal.
67
Appendix 1b. Time spent using the model evacuation plan with observations
No. 1 September 6, The students followed the new 3 minutes and

2017 (Recess evacuation plan but many were just 22 seconds
PM) walking towards the sink. There was less
congestion in the football sink.
No. 2 September 7, Stair B became less congested and 3 minutes and

2017 (Recess occupants in the left wing arrived at the 25 seconds
PM) sinks with less time. Most students
passing the OLSW path were just
walking. Many Grade 7 students were
playing and laughing. There was
congestion in OLSW path due to the
DAT.
No. 3 September 9, Congestion due to the DAT was 3 minutes and

2017 (Recess addressed. Many students are still 20 seconds
AM) walking and playing. Students in the
sinks were arranged and organized.
No. 4 September 19, Still, a smooth flow can be observed in 3 minutes and
2017 (Recess Stair B. Occupants of the AVR criss- 28 seconds
AM) crossed with other evacuees causing
congestion. Many students were
walking.
No. 5 September 20, Only a few students can be observed 3 minutes and
2017 (Recess walking. The presence of DAT hasten 15 seconds
AM) the evacuation activity. There was less
congestion in both stairs A and B.
68
Appendix 2. The Applicability of the mathematical model based on the demands of

the school’s present egress system
Problems in the present egress system Quickest flow problem approaches

of Holy Cross College of Calinan
Does not follow a model-based plan Network modelling with mathematical

approaches
Arrangement and distribution of students Sink Capacity and Sink Assignment

to different evacuation areas.
Longer path taken by some students Shortest Path Algorithm
Long time completion of evacuation Objective function: T

activity min ∑∑txid (t)
t=0 iϵD
Congestion of students in the stairs Node Capacity
Misutilization of available arcs and Source Distribution and Node Congestion

pathways
69
Appendix 3a. Distances from Sources to Sinks
2nd floor rooms to Stair A Exit meters Stair A to T1 (m) 3rd floor rooms to Stair A
201 13.71 79.025 301
202 20.67 Stair B to T1 (m) 302
203 27.73 58.835 303
204 34.71 OLSW exit to T1 (m) 304
205 41.74 27.5 305
206 43.28 Stair B to T2 (m) 306
207 53.37 21.895 307
208 63.65 Stair A to T3 (m) 308
2nd floor rooms to Stair B Exit meters 44 3rd floor rooms to Stair B
201 61.51 Stair B to T3 (m) 301
202 45.125 43.7 302
203 47.59 303
204 43.06 304
205 36.05 305
206 29.08 306
207 20.98 307
208 13.22 308
meters Stair A to T1 (m) 3rd floor rooms to Stair A meters 4th floor rooms to stair A meters
1 13.71 79.025 301 26.125 401 82.64
2 20.67 Stair B to T1 (m) 302 33.105 402 89.53
3 27.73 58.835 303 40.165 4th floor rooms to stair B meters
4 34.71 OLSW exit to T1 (m) 304 47.145 401 84.885
5 41.74 27.5 305 54.175 401 91.775
6 43.28 Stair B to T2 (m) 306 55.715
7 53.37 21.895 307 65.805
8 63.65 Stair A to T3 (m) 308 76.085
meters 44 3rd floor rooms to Stair B meters
1 61.51 Stair B to T3 (m) 301 74.91
2 45.125 43.7 302 84.845
3 47.59 303 77.805
4 43.06 304 72.785
5 36.05 305 66.225
6 29.08 306 61.245
7 20.98 307 52.365
8 13.22 308 44.365
70
Appendix 3b. Area Measurements of the Arcs
STAIR B AREA length width

stair 1 3.48 1.7
flat 1 3.73 1.67
stair 2 2.23 1.71
flat 2 3.46 2.41
stair 3 3.17 1.73
flat 3 3.78 1.66
stair 4 1.91 1.76
flat 4 3.78 2.65
stair 5 2.21 1.72
flat 5 3.86 1.66
stair 6 3.05 1.84
flat 6 3.69 2.49
ground floor flat 2.29 4.11
STAIR A AREA length width

stair 1 3.46 1.52
flat 1 3.27 1.44
stair 2 2.21 1.54
flat 2 3.16 2.44
stair 3 3.14 1.52
flat 3 3.28 1.41
stair 4 1.91 1.54
flat 4 3.3 2.47
stair 5 3.15 1.51
flat 5 3.28 1.41
stair 6 1.93 1.52
flat 6 3.28 3.39
ground floor flat 1.95
71
Appendix 4. Costs of Different Arcs in Time Units
Arc Cost
(A, B) 3 time units
(B, C) 3 time units
(C, I) 7 time units
(I, t1) 5 time units
(E, F) 3 time units
(D, E) 3 time units
(E, G) 4 time units
(F, t3) 6 time units
(F, H) 6 time units
(H, t1) 6 time units
(G, t1) 5 time units

72
Appendix 5. Capacity of different arcs relative to flow size
Arc Capacity
(A, B) 2 flows
(B, C) 2 flows
(C, t2) 3 flows
(C, I) 5 flows
(I, t1) 4 flows
(E, F) 2 flows
(D, E) 2 flows
(E, G) 3 flows
(F, t3) 5 flows
(F, H) 5 flows
(H, t1) 5 flows
(G, t1) 3 flows
(C, t3) 5 flows

73
Appendix 6. Capacity of different sinks based on the average occupancy
Sink Area Safe Zone Capacity
Football Field (t1) 1,158.4485 m2 891 m2 891 persons
Front of Guidance (t2) 524.8615 m2 364 m2 364 persons
Pieta (t3) 317.4233 m2 143 m2 143 persons
Grade school grounds (t4) 1,152.299 m2 820 m2 820 persons

74
Appendix 7. Distribution of Flows at each Time Interval
ARC TIME INTERVAL
A-B 0-29.09s 29.09 – 58.18 s 58.18 – 87.2 s
Arc TIME INTERVAL
B-C 0 – 29.09 s 29.09 – 58.18 s 58.18 – 87.2 s
2 3
Arc TIME INTERVAL
E-F 0 – 28 s 28 – 56 s 56 – 84 s
ARC TIME INTERVAL
E-G 0 – 45.74 s 45.74 – 91.48 s 91.48 s – 137.22 s
2 1
ARC TIME INTERVAL
G-T1 0 – 49.6 s 49.6 – 99.2 s 99.2 – 148.8 s
3 3
ARC TIME INTERVAL
C-T2 0 – 42.48 s 42.48 – 84.96 s 84.96 – 127.44 s
2 4
ARC TIME INTERVAL
C-I 0 – 69 s 69 – 138 s 138 – 207 s
ARC TIME INTERVAL
F-H 0 – 66.52 s 66.52 – 133.04 s 133.04 – 199.56 s
ARC TIME INTERVAL
I-T1 0 - 52.32 s 52.32 – 104.64 s 104.64 – 156.96 s
ARC TIME INTERVAL
H-T1 0 - 66.52 s 66.52 – 133.04 s 133.04 – 199.56 s
4
75
Appendix 8. Shortest Path Algorithm

76
Appendix 9. Arrangement of Sources in the Evacuation Areas

77
Appendix 10. Arc costs and Capacities

78
Appendix 11. Calculation of Sink Capacities
33.34 m
98 o 91o
130.5062 m2 7.38 m
6.75 m
114o
174.5667 m2 89o 9.42 m
45.2616 m2 80o
20.7802 m2
14.68 m 74.2890 m2 4.48 m
o
170o 29.1259 m 125
79.4578 m2
13.32 m
17.83 m 138o
79
Appendix 12. Distribution of Flow Units at Each Time Interval

80
Appendix 13a. Route Assignment based on QFP
81
Appendix 13b. Route Assignment without the Model

82
Appendix 14. Statistical Analysis in R

83
Appendix 15. Boxplot Function

84
Appendix 16a. Formulated Evacuation Plan Using the Model in PM Building

classrooms
85
86
87
Appendix 16b. Formulated Evacuation Plan Using the Model in the Different
Offices
88
89
Appendix 17. Earthquake drills using the non-model-based evacuation plan

90
Appendix 18. Earthquake drills using the model-based evacuation plan

91
Appendix 19. Measuring of School Dimensions

92
Appendix 20. Implementation of the Model

93
Appendix 21a. Letter of Permission to the School President

94
Appendix 21b. Letters of Permission to the School Principal

95
Appendix 21c. Letter of Permission to the Dean of College

96
Appendix 21d. Excuse Letter

Quickest Flow

Uploaded by

Copyright:

Available Formats

You might also like

Quickest Flow

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Quickest Flow

Uploaded by

Copyright:

Available Formats

HOLY CROSS COLLEGE OF CALINAN

CALINAN, DAVAO CITY

THE EFFECTIVENESS OF QUICKEST FLOW MODEL

Ong, Tian Ruy

THE EFFECTIVENESS OF QUICKEST FLOW MODEL

In Partial Fulfillment of the Requirements

Ong, Tian Ruy

In partial fulfillment of the requirements in Practical Research II, this study

entitled THE EFFECTIVENESS OF QUICKEST FLOW MODEL IN LESSENING

THE EVACUATION TIME IN HOLY CROSS COLLEGE OF CALINAN,

recommended for oral examination, approval and acceptance.

MELINA C. GONZALES, MAST, MES

ENG. MARLOU O. BIBANGCO MA. CORAZON C. SUNGA, M.A.

RIZALITO H. PAGA, M.A.

Accepted in partial fulfillment of the requirements in Senior High School.

Date of Oral Examination: October 5, 2017

MA. CORAZON C. SUNGA, M.A.

for allowing the researchers to conduct the study;

implementation of the study;

researchers facilitate the flow of the students during evacuation;

researchers address different setbacks and problems encountered in the study;

throughout their study;

improve their study;

strength, wisdom, and enlightenment to finish the study. Thank You.

In this research study, an earthquake evacuation strategy was developed through

as well as the possible number of casualties whenever an earthquake disaster occurs.

spent by the evacuees is significantly lower by up to 45 seconds when the model-based

the proposed strategy be employed as the school’s new evacuation plan.

Keywords: time-efficient evacuation, Quickest Flow Model, routing strategy, shortest

path algorithm, non-model-based evacuation plan

Chapter 3 RESULTS AND DISCUSSION

Chapter 4 CONCLUSION AND RECOMMENDATIONS

Appendix 2: The Applicability of the mathematical model 68

based on the demands of the school’s present

Appendix 3a: Distances from Sources to Sinks 69

Appendix 3b: Area Measurements of the Arcs 70

Appendix 4: Cost of Different Arcs in Time Units 71

Appendix 5: Capacity of different arcs relative to flow size 72

Appendix 6: Capacity of different sinks based on the average 73

Appendix 7: Distribution of Flows at each Time Interval 74

Appendix 8: Shortest Path Algorithm 75

Appendix 9: Arrangement of Sources in the Evacuation Areas 76

Appendix 10: Arc Costs and Capacities 77

Appendix 11: Calculation of Sink Capacities 78

Appendix 12: Distribution of Flow Units at Each Time Interval 79

Appendix 13a: Route Assignment Based on QFP 80

Appendix 13b: Route Assignment without the Model 81

Appendix 14: Statistical Analysis in R 82

Appendix 15: Boxplot Function 83

Appendix 16a: Formulated Evacuation Plan Using the Model in 84

Appendix 16b: Formulated Evacuation Plan Using the Model in 87

Appendix 17: Earthquake Drills Using the Non-Model-Based 88

Appendix 18: Earthquake Drills Using the Model-Based 90

Appendix 19: Measuring of School Dimensions 91

Appendix 20: Implementation of the Model 92

Appendix 21a: Letter of Permission to the School President 93

Appendix 21b: Letters of Permission to the School Principal 94

Appendix 21c: Letter of Permission to the Dean of College 95