Download as pdf or txt
Download as pdf or txt
You are on page 1of 35

Group Decis Negot (2017) 26:523–546

DOI 10.1007/s10726-016-9512-8

A Voting Approach Applied to Preventive Maintenance


Management of a Water Supply System

Adiel T. de Almeida-Filho1 · Madson B. S. Monte1 ·


Danielle C. Morais1

Published online: 27 September 2016


© Springer Science+Business Media Dordrecht 2016

Abstract In developing countries, water distribution is a service that has limited


resources for its expansion and modernization. Besides this problem, some regions
suffer shortages of water. Thus, this resource is under-supplied to many communi-
ties. For this kind of situation, the planning of maintenance activities becomes even
more important in order to reduce downtime due to system failures. Generally, more
than one person may be involved in such kinds of decision-making processes. Thus,
methods to aid group decision-making play an important role in finding a solution
that captures different points of view. Therefore, this paper puts forward an approach
for analyzing the preferences of water supply maintenance managers and includes the
customer’s perspective in the decision process. The evaluation criteria are set in terms
of seeking to establish the optimal interval for preventive maintenance. The alterna-
tive chosen based on the rankings from individual decision-makers (DMs) represents
the best compromise from the perceptions of all DMs involved in this problem. The
proposed approach to aggregate the DMs’ preferences is well suited to the context of
maintenance management for a water supply system.

Keywords Water supply · Preventive maintenance · Group decision-making ·


Quartile based voting

B Adiel T. de Almeida-Filho
adieltaf@cdsid.org.br
Madson B. S. Monte
madsonmonte@gmail.com
Danielle C. Morais
dcmorais@cdsid.org.br

1 Management Engineering Department, Universidade Federal de Pernambuco, Caixa Postal 7471,


Recife, PE 50.630-971, Brazil

123
524 A. T. de Almeida-Filho et al.

1 Introduction

Some water supply systems are unable to supply all consumers simultaneously, above
all in underdeveloped regions. This is primarily because there is no pairing between
urban growth and the expansion and modernization of distribution systems. Thus,
some areas have limitations related to water availability. Generally, there is a rationing
scheme, where each area receives water only in line with a specific schedule. Thus, in
such regions, the population is very adversely affected whenever the water supply is
interrupted.
The study presented in this paper has these characteristics and is located in a low-
income poor area, in the North Zone of the City of Recife, in North-East Brazil.
The population is supplied from a well which injects water directly into the network.
The water supply system (WSS) is managed by a company, the maintenance sector of
which has adopted a preventive maintenance policy, determined by the senior manager,
for the pump of this well. But currently there are two managers who are responsible
for all activities relating to well systems and they must agree with each other on
the maintenance policies to be adopted. They hold the mutual view that the supply
company should prioritize the benefit to the population, but they have to make trade-
offs with regard to operational criteria and sometimes disagree about how to measure
these attributes.
The study was conducted on a well which is isolated from other systems in the area,
so as to avoid the study being affected by external influences. Similarly, the criteria used
in planning how best to maintain this well are very specific, thereby requiring decision
makers to think more carefully about this environment before making remarks. A
range of criteria is needed so as to take account of the point of view of each decision-
maker (DM). The DMs will attach different degrees of importance to these criteria.
In addition, while managers recognize that consumers are the focus of the operation,
managers claim that consumers have never been considered as having a role in the
decision-making process when maintenance activities are being planned.
In maintenance engineering, different models can be proposed for evaluating the
characteristics of a system. Also, when properly applied, a model serves as a tool
for choosing the maintenance policy to be adopted. However, the choice of criteria
to be optimized depends on the DM and this choice can lead to a conflict since the
policy that optimizes one criterion does not necessarily optimize all others. The use
of multi-criteria decision-making (MCDM) models is an important tool for solving
this type of problem. In the case studied, the problem becomes more complex since
the solution must reconcile the views of both managers and consumers. Hence it is
important to apply a procedure that aggregates preferences.
Based on the foregoing, a case study case is developed which uses a voting proce-
dure to find an alternative that best reconciles the preferences of the DMs involved.
Besides the company’s two managers, a third party to this situation is added, namely
the consumers, who desire the criterion of availability to be met at the maximum fea-
sible level. This entails including a consumer as a DM who will express consumers’
expectations. This DM’s view is then used as a way to compare the result of aggregat-
ing the other DMs’ preferences, thereby enabling a better understanding of the impact

123
A Voting Approach Applied to Preventive Maintenance… 525

of decisions on supplying water from both the consumers’ and the company’s point
of view, with or without customer participation.
A differential of this proposal is the ability to describe all the criteria of DMs in
terms of a single variable. This variable is the interval between preventive maintenance,
which depends intrinsically on the condition of equipment and therefore the failure
behavior is analogous to a Weibull probability density function (Weibull 1951). This
ensures that each alternative represents a package of consequences from the point
of view of each criterion considered. A combination of reliability engineering and
group decision or voting methods can indicate an alternative which is the best com-
promise between the different points of view, while taking the parameters of system
into account.
Morais and de Almeida (2010, 2012) proposed a voting procedure applied in the
context of water resources management. This method is also applied in this paper. One
of its advantages is that not only may any method be used to rank DMs’ alternatives
but different methods for different DMs may be used. Furthermore, the procedure
eliminates the influence of irrelevant alternatives, thereby making the result even more
solid.
The paper is organized as follows: Sect. 2 gives a short review of the literature,
the focus being on the use of group decision methods or voting procedures for trou-
bleshooting to do with the management of water resources. Section 3 describes the
case study by introducing the DMs, the modeling criteria, the voting procedure used to
aggregate the DMs’ preferences and the results of this application. Section 4 discusses
the results and the properties of the method and Sect. 5 summarizes the paper and
makes final remarks.

2 Literature Review

Besides being used in political elections, voting procedures can play an important
role in supporting multi-objective decisions for a group of DMs. Each participant
must indicate his/her preferences whatever the criteria considered or the method used
to obtain their individual results. In the literature, several voting methods have been
proposed, some of which are presented in this section.
The fundamentals of voting theory emerged in the eighteenth century in the work of
Jean-Charles de Borda (Straffin 1980). Borda counting was criticized by Condorcet,
who instead proposed a pairwise comparison method. In Borda counting, the choice
made is based on the score given to each alternative using the DM’s preferences
that should indicate an order of alternatives. A count is made and the alternative
with the largest sum of weights is the Borda winner (Nurmi 1983). The Condorcet
procedure, on the other hand, consists of an assessment based on pairwise comparisons
of alternatives. The winning alternative is the one that displays an advantage over all
others for the largest number of DMs and is called the Condorcet winner (Straffin
1980). From this point of view, the Borda count is not a Condorcet winner method,
but it will never choose a Condorcet loser (Straffin 1980).
There are several voting procedures such as the fairly common Plurality system,
the simplest way to measure the collective will, in which each voter can vote for

123
526 A. T. de Almeida-Filho et al.

one option or abstain. The alternative with the largest number of votes is the winner
(Smith 1973). However, when there is no one alternative with more than 50 % of the
votes, a second ballot is held in which votes are cast for one of the two most voted
for alternatives in the first round. This is called a plurality runoff. Copeland’s method
is based on pairwise comparison and a score function. The scores for each alternative
xi is determined by the difference between the number of alternatives that xi defeats
and those that beat it (Nurmi 1983). Hare’s procedure and Coombs procedure are
based on the ranking of alternatives. However, if there is no immediate winner, Hare
eliminates the alternative with the smallest number of first places and a recount of
the votes given to the new set of alternatives is conducted. This procedure is repeated
until an alternative is found which is placed first by the majority of voters. Coombs
eliminates the alternative most often placed last and, as in Hare, the votes are then
recounted. According to Nurmi, Coombs’s procedure can be interpreted as one that
elects the “least intolerable”. Nurmi (2014) presents a discussion on the vagueness
and ambiguous aspects inherent to proportional systems of representation for voting
procedures.
It is not appropriate, in this paper, to cite all voting methods but it should be noted
that just as important as the individual preferences of DMs is the voting system adopted
to aggregate them. This is also crucial to the final outcome of the decision-making
process. Each procedure has its own characteristics that must be analyzed before it is
considered for selection. The complexity of Social Choice Theory is the basis since this
theory studies how individual preferences are added to form a collective preference.
Therefore, social choice theory explains Arrow’s contributions (Arrow 1950).

2.1 Group Decision Within Water Resources Management

Decisions on water resources in general depend on more than one person. The oper-
ational level, the management of companies or the wants of consumers themselves
can be considered. The theme has therefore been extensively explored by considering
different situations and methods.
In 2001, Hämäläinen et al. developed a tool to support decisions about the manage-
ment problem of a lake-river system that took all stakeholders’ views into account.
This tool consists of an interactive process that covered problem structuring and cre-
ating a value tree, evaluating individual Pareto-optimality, and combining preferences
in a model called preference programming to reach a consensus and thereby to choose
a management policy.
Madani et al. (2014) developed a stochastic framework. The decision making under
uncertainty is studied in a case study of a river delta. Water from the river had long
been used for urban and agriculture consumption and now needed interventions to
conserve the ecosystem and to ensure the reliability of the water supply. Monte–
Carlo social choice making is the name of a proposed method which is used to help
a partially cooperative group. It is called partially cooperative because, even though
the parties did not agree to a Pareto solution, they agreed to use a voting procedure
to select an alternative. Also addressing large water systems, Halabi et al. (2012)
considered linguistic labels and a fuzzy ranking procedure that is able to include the

123
A Voting Approach Applied to Preventive Maintenance… 527

mean, the standard deviation, the fuzzy membership function and the frequency of
experts’ opinions for each alternative.
Lienert et al. (2014) structured the problem of planning the infrastructure for a
water and sewage system in Switzerland. The study included various stakeholders
and resulted in creating four more suitable scenarios. The authors used several face-
to-face interviews and two workshops to create pillar issues from which to define
objectives and attributes for developing alternatives and structuring the scenarios. In
this study, MAUT (Multi-Attribute Utility Theory) was used to explore the individual
preferences.
Cullen (1989) raises the issue of water quality. This is a criterion that can be consid-
ered in studies on the distribution of water. This parameter has a stochastic character,
just like those related to engineering maintenance. Cullen (1989) presents the tradeoff
(de Almeida et al. 2016) that should exist between different criteria in the management
of water supply by linking the cost and the likelihood of contamination by bacteria.
Moreover, he states that these issues should not be evaluated only by professionals,
but also include the population.

2.1.1 DM ggregation Using a Multicriteria Approach

Morais and de Almeida (2007) detected a problem of water losses mainly as a result
of leakages in the water network. They developed a strategy for managing leakages
and allocating the scarce financial resources in the best way. But this is a complex
task due to other problems inherent to water distribution and sanitation, as is the prob-
lem previously dealt with by Hämäläinen et al. (2001) involving different objectives
and conflicts of interest among stakeholders. To develop the strategy, Morais and de
Almeida (2007) based their study on PROMETHEE-GDSS (Preference Ranking Orga-
nization Method for Enrichment Evaluation—Group Decision Support System) and
PROMETHEE V to select the alternative which made best use of the resources avail-
able. A case study in Brazil, conducted by Morais et al. (2014), dealt more efficiently
with the problem of water losses. They sort the areas, thereby generating the priorities
that direct the maintenance efforts by using the SMAA-TRI (Stochastic Multicriteria
Acceptability Analysis) method. Morais et al. (2014) also unravel the complexity of
maintenance for water distribution systems and show how to deal with different issues,
depending on the opinion of each DM. In another case involving maintenance, Trojan
and Morais (2012) selected a group decision approach to prioritize maintenance and
rehabilitation alternatives in different locations, using ELECTRE II (Elimination and
Choice Expressing Reality) and the Copeland method.
The management of watersheds was the problematic of d’Angelo et al. (1998),
where five different voting procedures were applied to reach a solution. Within the
same thematic, Silva et al. (2010) used PROMETHEE II to obtain a ranking of alter-
natives and posterior aggregation, considering groups of criteria in economic, social
and environmental spheres. In problems like this, there are a considerable number of
third parties who suffer the effects of decision but do not participate in the decision
process or else their interests are considered to be less important. Because of this, the
authors pursued how to decentralize and reduce the influence of powerful members.
Zhao et al. (2013) also used PROMETHEE for selecting the potential solution for

123
528 A. T. de Almeida-Filho et al.

tackling degradation in a river. In this case, the experts receive weights in accordance
with the degree of their support after they presented their rankings of the alternatives
instead of assigning them a priori weight subjectively.
To help decision making in the supply of urban water, Roozbahani et al. (2012)
proposed conflict resolution by combining a method of the PROMETHEE family with
the Precedence Order of the Criteria (PPOC). The authors report that DM satisfaction
increased because of the results obtained. The advantages presented were: it was easy
to apply and flexible as it can include a wide range of participants who have different
perceptions, and also because it can aggregate the uncertainties about the levels of
importance between the criteria.
A model to address the offsetting effects of an additive aggregation of DMs’ prefer-
ences is applied by Daher and de Almeida (2012) to the issue of allocating investments
to improve water distribution systems and sewage treatment in Brazil. Aggregation
may produce undesirable results and therefore a veto concept combined with the util-
ity of alternatives is applied in order to penalize conflicting alternatives and reduce
disagreements in the additive model. The veto is a constituent part of the procedure
applied in this paper and is also introduced for the same purpose, namely to eliminate
conflicts in decision-making.
Sun et al. (2015) presented an ecological water compensation program with multi-
attribute decision-making. They introduced the concept of attitude indices in order to
avoid deviations in evaluations of the DMs and based on minimum deviation, each
one had its weights associated. The difficulty of this topic is that ecological water
compensation program is a dynamic and uncertain process.
Villaverde et al. (2015) put forward an application of the AHP (Analytical Hierarchy
Process) method to achieve a better understanding of the complexity of decisions
regarding the management of water services. The study discusses the discrepancies
in the criteria applied and the difference between the weights set by experts in the
decision on privatizing the management of water supply services.

2.1.2 DM Aggregation Using Voting Procedures

Already in 1973, Burke et al. (1973) showed that issues related to water resources could
be analyzed from the point of view of group decision. However, the problems were
solved as individual decisions. Since then, several studies about this field have been
published, all of which primarily analyze the results of voting when a plurality system is
used (Schoolmaster 1984; Mann et al. 2013; Hersch and Pelkowski 2014; Latinopoulos
2014). It is emphasized that in these papers the view of the population/consumer was
central to decision-making.
Christophe and Tina (2015) applied a median voter model to support decision for
land-use in a river basin. They found there was a very strong political component with
regard to defending agricultural activities in rural municipalities. The problem is the
impact on aquatic flora and fauna and of reducing water retention in riparian zones
and the floodplain. In this case, the conflict between preserving agricultural activities,
permitting local development and nature conservation is clear.
Models are also found in the literature which link AHP methodology to voting pro-
cedures. Srdjevic (2007) applied a decision support system to the water committee of

123
A Voting Approach Applied to Preventive Maintenance… 529

the San Francisco river in Brazil. The proposed method applies AHP to subgroups of
DMs and after conducting an aggregation using various social choice methods (Plu-
rality voting, Hare system, Borda count, pairwise comparison method and Approval
voting). Srdjevic and Srdjevic (2013) followed the same reasoning, and applied AHP
with the Borda count to irrigation and draining systems. And, to select a supplier of
irrigation equipment, Draginčič and Vranešević (2014) used AHP with two voting
procedures: approval voting and Borda count. There were five DMs.
Morais and de Almeida (2010, 2012) developed a model in order to choose an
alternative regarding the rehabilitation of a water network and controlling degradation
in a hydrographic basin, respectively. They proposed a method, based on the Borda
count, that eliminates irrelevant alternatives from each DM’s ranking of the alternatives
and penalizes the worst classified before making the choice. In 2015, Cunha et al.
(2015) applied the voting procedure, proposed by Morais and de Almeida (2010,
2012), to a water crisis in the richest region of Brazil. They showed that a voting
procedure is a good option for aggregating individual opinions, even divergent ones.
In this paper, the voting procedure of Morais and de Almeida (2012) is applied in
combination with maintenance engineering models to define a time interval between
preventive interventions. These models are used to characterize the criteria in terms of
a single decision variable. The maintenance policy should be defined for a well from
which water is drawn and distributed to a low-income urban region so the managers
of the water distribution company and the consumers are the main stakeholders of this
decision process. Each of them has his/her own criteria and importance levels but that
does not matter since the input of the voting procedure is just a complete order of
alternatives, regardless of the form in which this was obtained.

3 Case Study

The region studied is located in Recife, Brazil, in a predominantly residential area.


It consists of a low-income population of very high density, namely it serves 30,000
inhabitants/km2 , according to Recife (2016). The topography of this area which is
located approximately 7 km from the city center is hilly. In the periphery of the city,
there are deficiencies in the provision of public services and this includes in the supply
of water: the community constantly suffers from shortages of water. The population of
this study is homogeneous and the lack of water is a notable feature of living conditions
for this entire community which therefore makes common cause in demanding that
this situation be improved. Although there are rivers nearby, potable water for human
consumption cannot be drawn from them. In addition, projects to bring water from
sources that are more distant or to desalinate sea water depend on investments from
government and are not included in the scope of this paper.
The City of Recife is only partially served by a continuous water supply system. Due
to urban growth not having been planned and the water shortage, deep wells have had
to be drilled to supply residents with water. However, the flow rate of the underground
water is not high enough to supply all consumers simultaneously. Therefore, the region
was divided into sections and a schedule was created so that each section receives water
only during previously agreed scheduled periods.

123
STATE BOARD OF TECHNICAL EDUCATION
BIHAR

Registration No : 1741520115
Roll No : 411741520115

Following are the marks obtained by HARSH KUMAR


of Buddha Institute of Technology, Gaya
at Semester IV of Diploma in Civil Engineering
Examination 2022 (EVEN) held in the month of OCT, 2022
-----------------------------------------------------------------------------------------------
SUBJECT FULL MARKS PASS MARKS MARKS OBTAINED
NAME OF SUBJECTS CREDITS INT FIN TOTAL FIN TOTAL INT FIN TOTAL GRADE
-----------------------------------------------------------------------------------------------
THEORY PAPERS
HYDRAULICS 3 30 70 100 28 40 25 43 68 C
ADVANCE SURVEYING 3 30 70 100 28 40 26 49 75 B
THEORY OF STRUCTURE 3 30 70 100 28 40 25 39 64 C
BUILDING PLANNING AND DRAWING 3 30 70 100 28 40 26 33 59 D
TRANSPORTATION ENGINEERING 3 30 70 100 28 40 25 41 66 C
-----------------------------------------------------------------------------------------------
PRACTICAL PAPERS
HYDRAULICS LAB 1 15 35 050 - 20 13 30 43 A
ADVANCE SURVEYING LAB 1 15 35 050 - 20 14 30 44 A
-----------------------------------------------------------------------------------------------
TERM WORK PAPERS
THEORY OF STRUCTURE LAB (TW) 2 07 18 025 - 10 06 15 21 A
BUILDING PLANNING AND DRAWING(TW) 2 15 35 050 - 20 14 31 45 A+
TRANSPORTATION ENGINEERING LAB (TW) 1 07 18 025 - 10 06 15 21 A
COURSE AUTO CAD/STAAD. PRO / 2 15 35 050 - 20 14 30 44 A
OTHERS(TW)
-----------------------------------------------------------------------------------------------
GRAND TOTAL 24 750 550
-----------------------------------------------------------------------------------------------
SGPA 7.83
-----------------------------------------------------------------------------------------------
REMARKS : First Class

Result Published On: 08/02/2023


530 A. T. de Almeida-Filho et al.

Fig. 1 Local grid

Figure 1 gives a diagram of the water grid in the area studied. This has five artesian
wells but the data provided in this study refers to Well 1. The total population served
is about 30,000 inhabitants. On the right, labelled as ‘RES’, there is the reservoir and
pipeline from another system, with more capacity but this is not linked up to the entire
water network in this area. Note that the areas supplied by Wells 1 and 2, as well as
those supplied by Well 5, are totally independent of the other areas which have their
own wells. This is a scheme that occupies a small portion of the entire system, which
consists of about 200 wells. Thus, it is clear that the scale of the problem is much higher
but the characteristics recorded in this paper are consistent only with the properties of
this area served by Wells 1–5. The results reported in this study cannot be extended to
other areas because each area has its own different characteristics both in relation to
the parameters of supply and as to the number of consumers supplied. For example,
there are different levels of rationing and population density and these factors have a
direct impact on the supply scheme and consequently on the planning of maintenance
activities. In addition, factors such as family incomes and the educational level of adult
residents can lead the population to re-assessing the quality of service that it expects.
Nevertheless, there are no obvious obstacles to applying the methodology, since it can
take any and all special features into account.

3.1 Water Supply Company and the Local Context of the Supply Service

The water supply company, COMPESA, is a mixed capital society, in which the
government of the state of Pernambuco is the majority shareholder. A contract was
entered into between the company and the state government that public services of
water supply and sanitation should be conducted exclusively by COMPESA under an
associate management regime between the two parties. The contract stipulates that the
company shall provide the distribution service in a regular, continuous and efficient
manner. As to efficiency, the service must be performed at the lowest possible cost,
in line with the standards set by regulatory agencies. Therefore, such agencies, which
are part of the government, are also responsible for setting the fees charged for the
water distribution service.

123
िबहार सरकार
Government of Bihar
फॉम / Form - XIII

िनवास माण-प / Residence Certificate


जला / District : गया, अनुमड
ं ल / Sub-Division : गया सदर, अंचल / Circle : मानपुर

माण-प सं या : BRCCO/2023/7253693 िदनांक : 28/07/2023

मािणत िकया जाता है िक हष कुमार (Harsh Kumar), िपता (Father) शैले कुमार (Shailendra Kumar), माता
(Mother) आशा देवी (Asha Devi), ाम / मोह ा - कु हार टोली, वाड सं या -50, डाकघर - बुिनयादगंज, िपनकोड - 823003, थाना -
मु फ सल थाना, खंड - मानपुर, अनुमड
ं ल - गया सदर, जला - गया, रा य - िबहार के थायी िनवासी ह |

थान : मानपुर
िदनांक : 28/07/2023 dgSign2031LocnDyna

(ह ता र राज व अ धकारी / Signature Revenue Officer )

QR Code क जाँच https://serviceonline.bihar.gov.in पोटल एवं Play Store पर उपल ध ServicePlus Mobile App से कर।
नोट: यह द तावेज DigiLocker पर भी उपल ध है।

Reference No: BRCCO/2023/7253693 To View: https://serviceonline.bihar.gov.in/officials//t/NTkKo/A216B111 Token No: A216B111


A Voting Approach Applied to Preventive Maintenance… 531

Also according to the public contract, the only reasons that make it legal for the water
delivery service to be interrupted are technical reasons or safety of the installations;
improper handling of distribution networks and/or of equipment that are company-
owned; consumer debt with the company or a force majeure event. Such force majeure
events should be justified with the regulatory company and include the regularly recur-
ring problems in the region such as water scarcity and system failure.
Access to clean water was not defined as a fundamental right in Brazil’s 1988 Con-
stitution (Brasil 2016a). However, it does regard seeking to contribute to human good
health as a fundamental right and it is known that access to clean water contributes
to healthy living. So, there is an indirect recognition of the need for access to clean
water and for this reason, local and Federal government are associated with distrib-
ution companies. Most other states in the country have set up companies similar to
COMPESA for which the state government of Pernambuco State holds more than 50 %
of the share capital and is therefore the main shareholder. Brazilian Law 9433 was
approved in 1997 (Brasil 2016b), 13 years before the United Nations considered water
as an essential human right, namely by UN resolution 64/292 of 2010 (UN 2016).
Despite Law 9433/1997 being on the statute book, there are still many locations in
Brazil today that do not have a public supply of water. Such locations depend on it
being possible to store rainfall water, and these include those in arid regions.
The State Government, together with the Water Company, established a flat rate tar-
iff so as to assist low-income families, who pay just a symbolic value for water service.
Furthermore, most of the local population fit the criteria defined by the government
and which need to be met to join this social program. Given the context of the Water
Supply Company, service improvements (which imply that costs will be raised) such
as expanding the service, rehabilitating sections of the grid and maintenance actions
do not lead to the tariff being raised.

3.2 Details of the Maintenance Problem and the Decision-Makers

A water supply system consists of several devices, but in the case of direct supply,
it is simpler to use wells. These are usually more than 100 m deep. The pump and
the rotor are the main and most expensive equipment, which is installed downhole.
Besides these items, the system consists of pipes and valves. The pumps operates con-
tinuously and, like any mechanical equipment, they are subject to breakage. However,
the inconvenience that would arise if the well pump were to break down is very large,
and therefore there is a need to reduce the likelihood of this occurring by planning
preventive maintenance actions.
Until 2012, the maintenance policy was based only on corrective maintenance,
which was performed only when the well ceased to supply water because the electric
engine or pump, which are submerged in the well, broke down. When this occurs, it
takes a lot of time and resources to remove them and the population is left without water
for a long period. Preventive maintenance reduces the likelihood of break-downs and
routinely requires less time and costs less than providing an emergency service. The
scheduled activities are a well-designed set of electrical, mechanical and plumbing
services aimed at re-establishing well conditions to an as-new state.

123
532 A. T. de Almeida-Filho et al.

However, a tradeoff often has to be considered when defining the frequency at


which the preventive maintenance will occur: if the interval between these activities
is very long, the high probability of a failure occurring remains. On the other hand, a
very short interval between preventive maintenance actions entails constant stoppages
and a high cost too. An ideal or optimum interval for preventive maintenance can
be found when only one criterion is considered but this does not correspond to this
situation. Besides the problem has a multi-criteria feature: the decision involves more
than one person within the organization, each with their own criteria. In the case of
maintenance sector, two engineers (DMs) are responsible for managing the activities
The maintenance managers agreed that looking in detail at each well can reveal
particularities that require consideration of different criteria. In addition, each DM
may have a different point of view and keep in mind different criteria with different
levels of importance. The quality of local water, the number of people supplied, the
level and depth of the well and the type of pump installed are examples of issues that
managers raise when planning preventive maintenance activities. Particularly in the
area selected, the criticality arising from the lack of water is very great and this leads
managers to seek to improve the system by using reliability parameters. The following
is a brief description of each part of the decision-making process.

• DM1 the first DM is a senior maintenance engineer of the water company and his
opinion is well respected within the company. It was he who proposed the change
in the maintenance policy, when it ceased to act only correctively on the network
equipment to act in order to prevent breakage. He tries to proceed proactively on
the system in order to avoid further costs and improve the distribution system. He
has a very simple and conservative view, considering two criteria for evaluation of
the consequences of his decisions when planning the activities of the maintenance
sector: the cost of maintenance activities and the availability of equipment for
water supply.
• DM2 He is also a maintenance engineer of the water supply company, but younger
than DM1. The DM2 perspective is different since his duties involve him in dealing
on a day-to-day basis base with the public. This requires him to have more contact
with the population that is supplied by water system. Therefore, he considers
different ways to measure the functioning of the system and to seek improvements.
For him, costs must always be tied to decisions, but as a service provider, he believes
that the priority should be to increase the reliability of the system and to reduce
the frequency at which the system fails as measured by SAIFI (System Average
Interruption Frequency Index).
• DM3 the third DM represents the local population that is supplied by the water
system. He has little technical knowledge of how the maintenance system operates.
However, he would like to have a service that is as good as possible, considering
the constraints of the system. Thus for DM3 the only aspect that should matter
in this decision is the availability of the water system, thus reducing shortages
in the supply of water. This is quite reasonable because, since the region suffers
constant rationing, a few more hours of water being supplied would be a great
benefit, thereby making residents’ domestic routines less uncomfortable. There
is no tradeoff related to the population when discussing service improvements

123
A Voting Approach Applied to Preventive Maintenance… 533

versus tariffs increase. Thus, it can be easily assumed that all users are ratio-
nal. Thus, they all have non-decreasing monotonic preferences regarding service
improvements (service availability). Therefore, any service improvement would
be preferred, although it would be true that the population’s perceptions on ser-
vice improvements are not homogeneous. This justifies why DM3 represents the
aggregated view of population. He may marginally evaluate service availability
heterogeneously, although he will evaluate optimal availability homogeneously.
Supply problems affect all residents of the study area similarly, in addition to which
the differences in the profiles of the residences with regard to levels of income and
consumption patterns are minimal. So, this information supports the proposition that
the local population is homogeneous and that there is convergence to a common
opinion about the supply of water as set out by DM3.
The DMs have different profiles and different views. It is not an easy task to assign
weights to the respective DMs, as these values reflect the degree of importance of
each and many people do not feel comfortable performing this type of assignment.
DM1’s opinion is highly regarded in terms of importance during the decision-making
process. However, despite his higher ranking in the company’s hierarchy, DM1’s posi-
tion throughout this process was always to be in favor of the notion that that all views
about the problem should have the same degree of importance, thereby obviating fur-
ther questioning on how the roles of each DM whether internally within the company
or externally had been defined and prioritized.
On the other hand, for such kinds of decisions, only technical aspects are usually
taken into account, which normally are concerned with a cost/benefit evaluation. Given
the characteristics of this community and the frequent shortages of water, a different
approach has been considered, which enables a higher level in the hierarchy (DM1),
local management (DM2) and the community’s view (DM3) all being included in
this decision-making problem. Moreover, note that graduate engineers do not usually
reside in areas like the one considered in this study and therefore they may well
not perceive as much as those who do reside in them how badly water shortages
affect local residents. Given the lack of knowledge about such kinds of decision, once
the community starts to participate, even though this does not give it a majority of
votes, the fact of its being included and its opinions listened to within the decision
process makes it satisfied that their opinions have been considered. Furthermore, all
stakeholders regard the community’s participation as an enhancement. This is not
limited to including the views of the population in the decision-making process, but
clarifies for all involved what the limitations of the local water supply system are in
terms of maximum availability or minimum cost.

3.3 Proposed Model

3.3.1 Criteria Modeling

Assuming a preventive maintenance policy based on the machine age, interventions


are performed on the equipment after τ time units in operation. This variable is defined
as the interval between preventive maintenance activities. It can be assumed that the

123
534 A. T. de Almeida-Filho et al.

equipment is subject to failure at any time, the time of occurrence of which is governed
by a probability distribution function with a Weibull-shape parameter (β) greater than
1 (Cassady and Kutanoglu 2005). Values of the parameter β greater than 1 indicate
that preventive maintenance can be effective in reducing the risk of total stoppage of
the machinery due to a fault.
Routinely, corrective interventions are of an emergency character, take longer and
cost much more. So, the consequence of defining a preventive maintenance interval
is to alter a set of stochastic variables: availability, system reliability, frequency of
downtimes, failure rate, cost of interventions, etc. These variables are common criteria
used in maintenance and reliability engineering (de Almeida et al. 2015a). Therefore,
the models presented below show how the criteria of interest relate only with the τ
variable. Thus, for DMs, the problem comes down to choosing a value of τ which
represents their expectations.
SAIFI is the acronym for the System Average Interruption Frequency Index, ini-
tially created by IEEE (2012) to measure the average number of interruptions to the
electricity supply in a year, but broadly applicable. This is directly proportional to the
failure rate λi and the proportion of consumers affected when there is a failure Si . In
the case studied, the sectors are divided so that each sector has approximately the same
number of customers. According to Li et al. (2003), the failure rate is the predominant
factor in determining SAIFI, which is obtained by (1).

SAIFIi = λi Si (1)

Working with equipment, the failure behavior of which can be described by a two-
parameter Weibull function (Weibull 1951) with a parameter shape β greater than one,
the failure rate λ steadily increases over time t and is written as shown in (2). η is
the scale parameter of Weibull function, also determined by fitting. In this situation,
alternatives with shorter intervals between preventive maintenance actions have lower
SAIFI levels, and also provide high levels of reliability.
 β−1
β t
λ (t) = (2)
η η

The direct relationship between SAIFI and the reliability R of the equipment using the
failure rate function is described by (3), where f (t) is the Weibull probability density
function.
f (t)
λ (t) = (3)
R (t)
The SAIFI measurement includes the population affected by the occurrence of a failure.
However, in many studies, the reliability of the equipment is considered as a criterion.
Reliability can be used as a criterion depending on the convenience since it is a proxy
for explaining the failure behavior of the equipment.
By definition, reliability is the probability that a piece of equipment does not fail
until a specific time t (de Almeida et al. 2015b). The reliability is written as in (4),
where β is the shape parameter and η is the scale parameter. A large value of reliability
is preferred.

123
A Voting Approach Applied to Preventive Maintenance… 535

 β
− ηt
R (t) = e (4)
Since reliability is the probability of the equipment not failing in 0 to t interval, the
calculation of [1 − R(t)] results in the probability of a failure occurring in the same
time interval. This information is obtained from the historical records of interven-
tions conducted on the equipment due to failures and adjusting the data by using an
exponential regression model.
Since the process is considered to be cyclical, the age of the equipment returns
to zero after each maintenance intervention (Rausand and Høyland 2004). For water
distribution, the equipment works 24 h a day, stopping only for maintenance. Therefore,
there are three states: running, stopped for preventive maintenance or stopped for
corrective maintenance. Each maintenance activity demands an amount of time called
t p and tc , respectively. According to Jiang and Ji (2002), availability can be expressed
as (5). An alternative is better, the longer its availability.

∫τ0 R (t) dt
A (τ ) = (5)
∫τ0 R (t) dt + tc F (τ ) + t p R (τ )

Expected downtime is a common criterion also used to evaluate the quality of services.
Many managers prefer this measure because they understand it better (Rausand and
Høyland 2004). To calculate it, simply multiply A by the period of interest. Thus a
ranking based on availability is always the same as a downtime-based ranking.
Several papers state that it is unwise to use a cost measure disassociated from a time
measure (Beichelt 1976; Glasser 1969; Chareonsuk et al. 1997; de Almeida 2012).
Thus, the cost criterion is defined as the cost per cycle K , measured in $/hour units.
Using the same assumptions as in the previous section, the cost model is associated
with the age-based model for preventive maintenance (Barlow and Hunter 1960). The
assumption of a cyclical system implies that preventive maintenance must not be a
simple repair and even considers exchanging important parts of the system. Therefore,
the cost analysis must be a detailed study to include all aspects of maintenance in the
cost of preventive maintenance costs (C p ) and of corrective maintenance (Cc ). These
terms are considered in a cost per cycle model K (6) (Jiang and Ji 2002) and this cost
is considered to be good, the lower its value.

CC [1 − R (τ )] + C p R (τ )
K (τ ) = (6)
∫τ0 R (t) dt

The optimization of each criterion equation gives different results for τ , thereby char-
acterizing the problem as having conflicting goals even for only one DM. However,
when performing the optimization, the limits of the decision space are defined. All
alternatives must be within the interval defined by optimization and any alternative
outside this is considered to be dominated, because its performance is lower in all cri-
teria simultaneously. Thus, after setting the set of alternatives and quantifying them in
terms of the criteria, a decision support method for each DM can be applied, the result
of which is a ranking of alternatives. This is a necessary condition for aggregating the
group’s preferences, which requires a ranking as input.

123
536 A. T. de Almeida-Filho et al.

3.3.2 Obtaining the Individual Rankings

To prevent strategic voting, multi-criteria decision-making approaches were used to


obtain the DMs’ ranking. The voting procedure applied does not have restrictions on
how the inputs were obtained. All that is required is a complete order of alternatives.
Since there is no restriction in this regard, this information is obtained from the most
suitable method chosen according to the reasoning of each DM.
Thus, for DM1 the most suitable approach was Multi-attribute Utility Theory
(MAUT), which was applied in accordance with the procedures provided by Keeney
and Raiffa (1976). The reason for using this method is to identify a compensatory
rationing for DM1, who was assessed with regard to the tradeoff amongst criteria (de
Almeida et al. 2016; Keeney and Raiffa 1976). For DM1, increases in the cost are valid
given the return of the greater availability of the system for the population. Further-
more, this problem is a decision under risk. Utility independence between criteria was
verified, which allows the use of a multilinear multi-attribute utility function (Keeney
and Raiffa 1976). As the utility function is dependent on the preventive maintenance
interval, each alternative presents a utility level, which can be ordered to generate a
ranking of alternatives.
For DM2, the compensatory reasoning was not easy to perceive, so PROMETHEE
II (Brans and Mareschal 1984) was chosen to obtain the ranking for DM2. Simplicity
and ease of understanding were points that contributed to choosing PROMETHEE
II from among other outranking methods. In addition, as the problematic of this
method is ordering, it will provide a ranking which is the input of the voting pro-
cedure.
For DM3, as availability is the only relevant criterion, from (5) it is possible to
evaluate each alternative and order them. It is not easy for consumers to consider other
criteria in this decision context, since they are lay people in the technical aspects of
engineering maintenance and reliability.

3.3.3 Voting Procedure

The purpose of using the methodology applied in this paper is to select one alternative
based on rankings established from individual analysis of each decision maker. It was
developed by Morais and de Almeida (2012).
The input of this method is each DM’s ranking of the alternatives. These rankings
are divided into quartiles, where only the first and the last quartiles are analyzed.
The alternatives of the first group are strong candidates for the DMs to choose and
they receive a score in accordance with their position in the ranking. However, those
present within the last quartile are penalized. The indices applied in the model are called
the strength and weakness of the alternatives. This is a way to consider rejecting a
DM’s alternative and is therefore a differential of this method (Morais and de Almeida
2012). That is, even though the only purpose of this method is to choose an alternative,
information about the DMs’ rankings is used to enrich the process, thus balancing the
scores for each alternative.
The procedure consists of four steps. The first is to create a set of higher alter-
natives and lower alternatives. Therefore, each ranking from m DMs is divided into

123
A Voting Approach Applied to Preventive Maintenance… 537

Fig. 2 The first phase of the method

quartiles but only the first and fourth are used, representing the 25 % of alternatives
present in the top quartile position and the 75 % below this quartile, as shown in
Fig. 2. All alternatives cited in the first quartile must be considered in the set of the
highest-ranked alternatives, even if only one DM has voted for only one of them. The
other alternatives are disregarded. (7) is used to calculate the position in the ranking
which fits into the first quartile, where X must be rounded up and n is the number of
alternatives.
X = n/4 (7)
The lower quartile consists of alternatives that DMs voted below the position given
by (8). These alternatives are weaker and the function of the quartile is to penalize
those best place alternatives. Of the alternatives cited in the fourth quartile, only those
that are ranked highest must be considered, and comprise the set of alternatives that
should be penalized.
Y = (3n/4) + 1 (8)
The value of Y must be truncated.
In the second step, the strength (F) and the weakness ( f ) of alternatives should be
calculated as per the following equations. Consider

1, if the alternative i is in position j for the DM k
qikj =
0, otherwise

Thus, for the highest-ranked alternatives, calculate


m 
X
Fi = (X − j + 1) qikj ∀i, k∀ j = 1, . . . , X (9)
k=1 j=1

123
538 A. T. de Almeida-Filho et al.

And for those alternatives which must be penalized, calculate


m 
n
fi = ( j − Y + 1) qikj ∀i, k∀ j = Y, . . . , n (10)
k=1 j=y

Note that the highest-ranking alternatives earn a score in accordance with their position
in the first quartile and are penalized in accordance with their position in the fourth
quartile. The score and penalty are given in a similar way to the Borda method (Morais
and de Almeida 2012; Nurmi 1983). This procedure must be followed for all DMs’
rankings and the sum of points gained is computed, the result being the strength of
the alternative. In the same way, the total of penalties received must be summed to
calculate the weakness.
The third step considers the possibility of veto for each classified alternative. This
phase evaluates whether there is a high disagreement about selecting an alternative and
those classified as worst for most of the DMs are eliminated. The analysis is performed
by only comparing the strengths (Fi ) and weaknesses ( f i ) of the alternatives. For an
alternative i, if f i ≥ Fi , there is a strong opposition to this alternative and therefore it
must be eliminated.
The fourth and last phase consists of choosing the best alternative by considering
the differences of the strength and weakness indices (α), as in (11). The alternative
with the greatest α is the one chosen by the group.

αi = Fi − f i (11)

Note that the similarity to the Borda procedure includes there being a similar negative
property: the possibility of a strategic manipulation of the result. If one of the DMs in
the group wants to prevent an alternative being chosen so as to increase the chances
of his/her alternative being chosen, he has only to rank the undesired alternative last.
Thus it is more penalized and loses points when α calculus is performed on using
the difference of the indices (phase 4). In other words, this alternative is approved in
the first phase as a potential alternative but loses many points because of a coalition
against it by some DMS in the group. However this characteristic does not invalidate
its application in groups where it is expected that each DM sets his/her preferences
according to their own will. For this to be assured, the use of a multi-criteria method
to support obtaining the rankings is recommended.
To continue with the comparison with Borda, an advantage in favor of the method
of Morais and de Almeida (2012) is claimed. Their method reduces the influence of
irrelevant alternatives since it is applied only to 50 % of the alternatives (bottom and top
quartiles). The other alternatives have neither a positive nor a negative influence and
thus they are eliminated. This raises the consistency of results in terms of an Arrow
condition. In relation to other conditions stated by Arrow, Morais and de Almeida
(2012) ensure that their method allows all DMs to express their preferences and it
must account for all them. Furthermore, the preferences are revealed without other
DMs imposing and influencing them (non-dictatorship).

123
STATE BOARD OF TECHNICAL EDUCATION
BIHAR

Registration No : 1741520115
Roll No : 611741520115

Following are the marks obtained by HARSH KUMAR


of Buddha Institute of Technology, Gaya
at Semester VI of Diploma in Civil Engineering
Examination 2023 (EVEN) held in the month of SEP, 2023
-----------------------------------------------------------------------------------------------
SUBJECT FULL MARKS PASS MARKS MARKS OBTAINED
NAME OF SUBJECTS CREDITS INT FIN TOTAL FIN TOTAL INT FIN TOTAL GRADE
-----------------------------------------------------------------------------------------------
THEORY PAPERS
ENTREPRENEURSHIP AND START-UPS 3 30 70 100 28 40 28 37 65 C
PUBLIC HEALTH ENGINEERING 3 30 70 100 28 40 28 33 61 C
ADVANCED DESIGN OF STRUCTURES 3 30 70 100 28 40 28 37 65 C
TENDERING AND ACCOUNTS 3 30 70 100 28 40 28 39 67 C
INDIAN CONSTITUTION 2 30 70 100 28 40 27 43 70 B
-----------------------------------------------------------------------------------------------
PRACTICAL PAPERS
PUBLIC HEALTH ENGINEERING LAB 2 20 30 050 - 20 18 25 43 A
-----------------------------------------------------------------------------------------------
TERM WORK PAPERS
SEMINAR 2 15 35 050 - 20 13 31 44 A
MAJOR PROJECT 3 15 35 050 - 20 14 30 44 A
ARTIFICIAL INTELLIGENCE (ADVANCE) 1 20 30 050 - 20 18 26 44 A
TENDERING AND ACCOUNTS (TW) 2 15 35 050 - 20 14 31 45 A+
-----------------------------------------------------------------------------------------------
GRAND TOTAL 24 750 548
-----------------------------------------------------------------------------------------------
40% of SEMESTER - I TO V EXAMS 1500 1126
-----------------------------------------------------------------------------------------------
AGGREGATE 2250 1674
-----------------------------------------------------------------------------------------------
CGPA 8.19
-----------------------------------------------------------------------------------------------
REMARKS : First Class

Result Published On: 15/10/2023


540 A. T. de Almeida-Filho et al.

Table 2 The order of the


Ranking DM1 DM2 DM3
alternatives
1 A12 A5 A13
2 A11 A6 A12
3 A13 A7 A14
4 A10 A4 A15
5 A14 A8 A11
6 A9 A3 A10
7 A15 A9 A9
8 A8 A2 A8
9 A7 A10 A7
10 A6 A11 A6
11 A5 A1 A5
12 A4 A12 A4
13 A3 A13 A3
14 A2 A14 A2
15 A1 A15 A1

described in Keeney and Raiffa (1976). The alternatives are ordered as per their utility
level.

U = 72.9e(−2285(1−A)) + 12.94e−3.65K + 523.91e−(2285(1−A)+3.65K ) (12)

DM2 presented the following parameters: reliability criteria, SAIFI and cost with
weights 0.4; 0.4 and 0.2 respectively; U-shape function for all criteria and indifference
threshold of 0.1 to reliability and SAIFI and 0.2 to cost.
For the mono criterion decision of DM3, the alternatives were ordered as per their
expected level of availability, calculated from (5). Table 2 shows the rankings of the
three DMs.
Note that how DM1 and DM3 ordered the alternatives are similar. In part, this is
because the criterion of availability, which is common to these DMs and the scaling
constant attributed to it by DM1 in (12). On the other hand, this specific region has the
characteristics of a low-income population with a low level of education. Therefore,
their lesser understanding of technical aspects of the water system leads them to a
simplified evaluation based only on its availability.
Due to such local characteristics, it is not easy to generate an understanding of the
various ways for measuring the quality of service. Thus, availability turns out to be
the only criterion for evaluating the system from DM3’s point of view. In fact, for
DM3 there is a belief that the more investment is made in equipment maintenance,
the better the levels of availability that would be achieved. Although it may sound
contra-intuitive, maintenance engineering provides tools for defining the optimal inter-
val for preventive maintenance in order to establish a balance between interruptions

123
A Voting Approach Applied to Preventive Maintenance… 539

Table 1 Identification of
Alternative τ Availability Cost ($/hour) SAIFI
alternatives and consequences
A1 500 0.991989 1.7464 0.0533
A2 750 0.994567 1.2835 0.0914
A3 1000 0.995828 1.1020 0.1340
A4 1250 0.996557 1.0368 0.1803
A5 1500 0.997018 1.0321 0.2299
A6 1750 0.997325 1.0635 0.2822
A7 2000 0.997534 1.1186 0.3371
A8 2250 0.997678 1.1900 0.3942
A9 2500 0.997775 1.2733 0.4536
A10 2750 0.997839 1.3652 0.5149
A11 3000 0.997878 1.4635 0.5781
A12 3250 0.997897 1.5665 0.6431
A13 3500 0.997902 1.6560 0.6992
A14 3750 0.997894 1.7812 0.7780
A15 4000 0.997878 1.8908 0.8477

3.4 Results of the Application

The water supply company kindly provided the historical records of maintenance
already performed in the well, both preventive and emergency interventions. This
enabled learning what the behavior of faults was and adjusting the data to the Weibull
probability density function (Weibull 1951). The shape parameter β = 2.3306, greater
than one, means that the failure rate increases over time and that preventive interven-
tions can extend the useful life of the equipment. After defining the parameters of the
Weibull function, which represents the failure behavior of the equipment, throughout
(1), (5) and (6) it is possible to calculate the consequences of each alternative in terms
of the criteria adopted respectively by the DMs. Thus, the question now is to know
what preventive maintenance interval (τ ) should be applied.
The optimization of criteria functions provides contradictory values of τ , i.e. even
for a single DM if more than one criterion is taken into account, this choice may involve
a conflict. For instance, consider the criteria of availability and cost. When optimiz-
ing the cost involved for equipment maintenance, the optimal time interval between
preventive maintenance actions is about 1397 h, while to maximize availability, the
optimal time interval would be 3461 h.
As τ is a continuous variable, it has to be discretized so as to generate the set of
alternatives shown in Table 1, so, the alternatives have been considered within a 250-h
interval due to a scheduling convenience (water supply calendar) of up to 4000 h at
the request of the maintenance managers. This table also displays the magnitude of
the expected consequences of each alternative.
For DM1, (12) represents the multi-attribute utility function of DM1 (Monte and
Almeida-Filho 2016), where A and K are the criteria functions of availability (5) and
cost (6). The multi-attribute utility function was built using the traditional procedure

123
A Voting Approach Applied to Preventive Maintenance… 541

Table 3 Count of strengths and weaknesses

Alternative First quartile Last quartile


DM1 DM2 DM3 Strength DM1 DM2 DM3 Weakness

A4 0 1 0 1 1 0 1 2
A5 0 4 0 4 – – – –
A6 0 3 0 3 – – – –
A7 0 2 0 2 – – – –
A10 1 0 0 1 – – – –
A11 3 0 0 3 – – – –
A12 4 0 3 7 0 1 0 1
A13 2 0 4 6 0 2 0 2
A14 0 0 2 2 0 3 0 3
A15 0 0 1 1 0 4 0 4

due to equipment failure and interruptions due to preventive maintenance services,


as described in (5). So depending on the region of the availability curve (5), fewer
investments in preventive maintenance should be made in order to increase the avail-
ability of the supply. Thus, for the general understanding, availability was considered
as it provides a clear measure to evaluate the continuity of the service in terms of the
percentage of time that the system would be in operation.
For DM2, the best-placed alternatives are those that provide a more reliable system,
with a lower frequency of interruptions in water supply. Thus, the DMs see the concept
of better service delivery differently, which explains the disparity between how DM2
and DM3 order the alternatives.
Thereafter it is possible to start the voting procedure. There are fifteen alternatives.
From (7), there are positions which will be part of the first quartile and (8) gives the
number of alternatives that fit into the last quartile.
X = 15/4 = 3.75 → The top four positions are in the first quartile
Y = (3 ∗ 15/4) + 1 = 12.25 → From 12th position, the alternative is in the last
quartile
Table 3 shows all alternatives placed in the first quartile by all DMs, i.e. the highest-
ranked alternatives. (9) is applied to calculate the strength of these alternatives. The
Table also shows the alternatives in the fourth quartile and the results from calculating
the weaknesses from (10). Note that the alternatives A1, A2 and A3 do not appear in
the analysis of the last quartile as they were not mentioned as highest-ranked alterna-
tives.
The third phase of voting is the veto. The alternatives that have a weakness index
greater than their strength should be vetoed. Morais and de Almeida (2012) state that
in this situation there is a high opposition to the alternative being considered among
the best ones and it should be eliminated. This condition prevents the alternatives
A4, A14 and A15 from being considered as to determining their intensity α. When
these alternatives are removed, the remaining candidates are those with the greatest
potential to be chosen. Note that at this veto stage, a more restrictive condition can be

123
542 A. T. de Almeida-Filho et al.

Table 4 The choosing phase–α


Alternative F f Intensity α
calculus
A5 4 0 4
A6 3 0 3
A7 2 0 2
A10 1 0 1
A11 3 0 3
A12 7 1 6
A13 6 2 4

considered since this was in agreement with the DMs but this does not occur in this
study.
The fourth and last phase is that of making the choice. For all alternatives that have
not been eliminated, (11) provides the intensity of the alternative, thereby revealing
the one that should best represent the common view of this group of DMs. The result
is shown in Table 4.
The result of this process indicates alternative A12 as the winner. This is a satisfac-
tory result since A12 was placed first by DM1 and second by DM3. DM2 has a small
objection in relation to this choice. However, his dislike of it is not enough to remove
it from the top. On looking at Table 4, it is noticeable that alternatives A5 and A13
are also strong candidates. This is unsurprising as they were placed first by DM2 and
DM3, respectively.

4 Discussion

In comparison to results from other methods, in this situation, the applied method
is better than plurality because by this method, there is a tie between the three top
alternatives A12, A5 and A13. A12 is also found as the result of Borda’s count and
the Condorcet pairwise comparison.
The expected consequences of choosing A12 are as follows: there is a 20.13 %
probability that the well will require emergency maintenance before a preventive
intervention; a SAIFI of 0.6461 resulting from a failure rate of 1.61E-4 applied to
4000 people affected simultaneously in a failure; a maintenance cost estimated at
$13,720 per year funded by the supply company; and a high level of availability, with
approximately only 19 h of downtime in a year.
In addition, the participation of a representative of the population had an effect
on this decision. If DM3 had not been considered in this decision, the alternative
chosen according to the rankings from DM1 and DM2 would have been alternative A5.
This would have a lower performance precisely in the single criterion considered by
consumers. Thus, this shows how important it is for a representative of the population
to take part in the decision-making process if the quality of the public service provided
is to be improved.

123
STATE BOARD OF TECHNICAL EDUCATION
BIHAR

Registration No : 1741520115
Roll No : 311741520115

Following are the marks obtained by HARSH KUMAR


of Buddha Institute of Technology, Gaya
at Semester III of Diploma in Civil Engineering
Examination 2021 (ODD) held in the month of JUN, 2022
-----------------------------------------------------------------------------------------------
SUBJECT FULL MARKS PASS MARKS MARKS OBTAINED
NAME OF SUBJECTS CREDITS INT FIN TOTAL FIN TOTAL INT FIN TOTAL GRADE
-----------------------------------------------------------------------------------------------
THEORY PAPERS
BUILDING CONSTRUCTION AND 5 30 70 100 28 40 27 39-4 62 C
CONSTRUCTION MATERIALS
BASIC SURVEYING 3 30 70 100 28 40 25 38 63 C
MECHANICS OF MATERIALS 3 30 70 100 28 40 26 24+4 54 D
CONCRETE TECHNOLOGY 3 30 70 100 28 40 27 36 63 C
GEO TECHNICAL ENGINEERING 3 30 70 100 28 40 26 31 57 D
-----------------------------------------------------------------------------------------------
PRACTICAL PAPERS
BASIC SURVEYING LAB 1 15 35 050 - 20 13 26 39 B
BUILDING CONSTRUCTION AND 1 15 35 050 - 20 13 28 41 A
CONSTRUCTION MATERIALS LAB
WEB TECHNOLOGY LAB 1 07 18 025 - 10 06 16 22 A
-----------------------------------------------------------------------------------------------
TERM WORK PAPERS
MECHANICS OF MATERIALS LAB (TW) 1 07 18 025 - 10 06 15 21 A
CONCRETE TECHNOLOGY LAB (TW) 1 07 18 025 - 10 06 15 21 A
PYTHON 1 07 18 025 - 10 06 15 21 A
GEO TECHNICAL ENG. LAB (TW) 1 15 35 050 - 20 13 30 43 A
-----------------------------------------------------------------------------------------------
GRAND TOTAL 24 750 507
-----------------------------------------------------------------------------------------------
SGPA 7.29
-----------------------------------------------------------------------------------------------
REMARKS : First Class (U/R)

Result Published On: 12/08/2022


A Voting Approach Applied to Preventive Maintenance… 543

In this case study, the users’ preferences were considered, which guided the final
result of a policy that best benefits the population. However, it is pointed out that it is
difficult for DM3 to evaluate the capacity of the alternatives and consequences since he
may well have very limited knowledge about the system, maintenance engineering and
reliability. In this case, only one criterion was used which happens to have been one that
directly reflects their day-by-day living conditions DM3 always prefers an alternative
that provides greater availability but he does not understand how this impacts the
probability of failure and the consequent increase in the frequency of emergency
interventions in the well. According to Cullen (1989), the population needs to be
able to understand the system better if public participation in decisions about water
resources is to be effective.
Finally, as to the voting procedure, despite the problematic involving only three
DMs, the method is easy to use and can safely be extended to a large number of DMs.
When it is possible for the DMs to order the alternatives, the method is perfectly
adequate as long as the purpose of the group is to make a choice. However, as the
procedure does not exclude the possibility of manipulating the results, the use of
MCDM models are strongly recommended in order to obtain an individual’s real
preferences.
An interesting characteristic is that the method eliminates irrelevant alternatives.
These are located in the intermediate quartiles of the ordered alternatives that the DMs
define but do not evaluate as being either preferred or rejected, thereby focusing atten-
tion on the important ones. Other positive aspects of the voting procedure is the ease
with which the analyst makes evaluations; its ability to deal with DMs individually,
thus avoiding problems to do with DMs’ agendas; it is easy to include/exclude a DM,
if necessary; the possibility of incorporating the veto, if DMs so desire and agree; and
the possibility that although it was deemed DMs had equal levels of importance in this
context, in the procedure it is possible to assign weights to them.

5 Final Remarks

The application of decision support methods always makes a DM reflect better on the
problematic of reevaluating their criteria, their degree of importance and how they
relate to each other. An advantage of the proposed method is the freedom that one
has to extract the DM’s preferences, which only requires an ordering of alternatives
as input.
In this paper, reliability engineering models are used to describe the criteria set by
DMs. Therefore, it is possible to define values for the stochastic variables, thereby
facilitating how to interpret the characteristics of the system. Two imminent benefits
stand out: although it has worked with three criteria, all are functions of a single
decision variable and the modeling of criteria makes it even more difficult to opt for
strategic voting, if that were of interest to some DM.
In relation to the result, it is considered that this was quite satisfactory and revealed
what the group wished. So, the voting procedure was appropriate for aggregating the
DMs’ preferences. Furthermore, the inclusion of a DM representing the consumers
was extremely important because it directed the result to an alternative with better than

123
Government of Bihar (Candidate Copy)
Department of Science & Technology .

State Board of Technical Education Bihar, Patna .

ADMIT CARD .

Registration No. 1741520115 Board Roll No. 521741520115

Name HARSH KUMAR

2022 (ODD)
Semester V Year of Examination Suppl_Comp

Examination Center Government Polytechnic, Gaya

Institute Buddha Institute of Technology, Gaya

Branch Civil Engineering

Type Subject Code Subject Name Date Time - Slot


T 2015501 Design of steel and R.C.C Structure 19-Jun-2023 02:00 PM - 05:00 PM
.
.
.
.
.
.
.
.
.
.
.
.
.

Controller of Examination

Important Instruction
Candidates should, in addition to their Admit Card, bring: at least one photo identity proof in ORIGINAL with clear photograph (e.g. AADHAR
1. Card, Driving License, College ID, Voter ID Card, PAN Card, etc. If a candidate fails to bring the above documents, he/she would not be allowed
to enter the examination venue.
2. Candidates must report to the examination center at least one hour before the commencement of the examination.
3. Candidates are not allowed to enter the examination premises after the commencement of the examination.
4. Candidates are not allowed to leave the examination hall before the end of examination.
Prohibited items such as books, paper chits, magazines, electronic gadgets (Mobile phones, Bluetooth devices, Head phones, Scanner, Smart
5. Watch, Storage devices etc) are STRICTLY NOT ALLOWED in the Examination Hall. If any such item is found in the possession of a candidate
in the examination hall, his/her candidature is liable to be cancelled and legal/criminal proceedings shall be initiated against him/her. He/she would
also be liable to be debarred from appearing in future examinations of the SBTE for a period of 3 years.
6. The candidates are advised to avoid talking to other candidates or indulging in any unfair means.
7. Candidature of candidates is purely provisional. It is advised that candidates should satisfy themselves that they fulfill all the eligibility conditions.
If at any stage it is found that a candidate does not fulfill any of the eligibility conditions, his/her candidature shall be cancelled.
8. Candidates shall be permitted to appear in the exam only at the venue as specified in the Admit Card.
9. CANDIDATES MUST STRICTLY ABIDE BY THE INSTRUCTIONS GIVEN BY THE EXAM FUNCTIONARIES (CENTRE
SUPERINTENDENT/ INVIGILATOR, ETC).
10. If any candidate is found obstructing the conduct of the examination or creating disturbances at the examination venue, his/her candidature shall be
summarily cancelled.
11. Ensure that you have signed the attendance sheet for the examination separately circulated by the invigilator.
12. In case of any doubt or any clarification, the candidates are advised to contact with Centre Superintendent.
544 A. T. de Almeida-Filho et al.

expected consequences for the population, because of the increase in the availability
of the supply of water.
Participatory budgeting has been greatly encouraged in recent years and is an issue
that many communities raise since it may have quite positive results regarding the
efficiency of public expenditures. Although this practice is closely linked to the main
public policies of the politics of populism prevalent in many South American countries,
the key aspect is that when ordinary people are included in decision-making proce-
dures, this sees to it that they are informed about technical aspects of possible decisions
and better understand the need for efficiency when incurring public expenditures.
The particular case addressed in this paper shows an example of the point raised
by Elinor Ostrom and her coauthors, who argue that decision-making procedures
that do not take into account users’ opinions are usually inefficient (Ostrom 1990).
Thus, technical details are expressed in terms that are based on users’ opinion on
the subject, and therefore prioritize the technical aspects related to service level. A
feasible compromise of all technical aspects (including service level) leads to choosing
an efficient solution with respect to all the criteria considered. Thus, this paper may
be considered an example of “Doing Good with Good OR”.

Acknowledgements The authors are grateful for the support from CAPES, COMPESA (Sanitation Com-
pany of Pernambuco) and for the partial support of CNPq.

Conflict of interest The authors have no conflicts of interest to declare.

References
Arrow K (1950) A difficulty in the concept of social welfare. J Polit Econ 58:328–346
Barlow R, Hunter L (1960) Optimum preventive maintenance policies. Oper Res 8(1):90–100. doi:10.1287/
opre.8.1.90
Beichelt F (1976) A general preventive maintenance policy. Mathematische Operationsforschung und Sta-
tistik 7:927–932
Brans JP, Mareschal B (1984) PROMETHEE: a new family of outranking methods in multicriteria analysis.
Oper Res 84:408–421
Brasil (2016a) Constituição da República Federativa do Brasil de 1988. Palácio do Planalto. http://www.
planalto.gov.br/ccivil_03/Constituicao/Constituicao.htm. Accessed 27 May 2016
Brasil (2016b) Lei N◦ 9,433 de 8 de Janeiro de 1997. Palácio do Planalto. http://www.planalto.gov.br/ccivil_
03/Leis/L9433.htm. Accessed 27 May 2016
Burke R, Heaney JP, Pyatt EE (1973) Water resources and social choices. J Am Water Resour Assoc
9(3):433–447. doi:10.1111/j.1752-1688.1973.tb01758.x
Cassady CR, Kutanoglu E (2005) Integrating preventive maintenance planning and production scheduling
for a single machine. IEEE Trans Reliab 54(2):304–309. doi:10.1109/TR.2005.845967
Chareonsuk C, Nagarur N, Tabucanon MT (1997) A multicriteria approach to the selection of preventive
maintenance intervals. Int J Prod Econ 49:55–64. doi:10.1016/S0925-5273(96)00113-2
Christophe B, Tina R (2015) Integrating water resource management and land-use planning at the rural-
urban interface: insights from a political economy approach. Water Resour Econ 9:45–59. doi:10.
1016/j.wre.2014.11.005
Cullen P (1989) Social choice, risk and determinism in water quality management. Hydrobiologia 176:1–5.
doi:10.1007/BF00026538
Cunha A, Morais DC (2015) Decision support model for participatory management of water resource.
In: Delibašić B, Hernández JE, Papathanasiou J, Dargam F, Zaraté P, Ribeiro R, Liu S, Linden I
(eds) Lecture notes in business information processing. Springer, Berlin, pp 85–97. doi:10.1007/
978-3-319-18533-0_8

123
STATE BOARD OF TECHNICAL EDUCATION
BIHAR

Registration No : 1741520115
Roll No : 511741520115

Following are the marks obtained by HARSH KUMAR


of Buddha Institute of Technology, Gaya
at Semester V of Diploma in Civil Engineering
Examination 2022 (ODD) held in the month of APR, 2023
-----------------------------------------------------------------------------------------------
SUBJECT FULL MARKS PASS MARKS MARKS OBTAINED
NAME OF SUBJECTS CREDITS INT FIN TOTAL FIN TOTAL INT FIN TOTAL GRADE
-----------------------------------------------------------------------------------------------
THEORY PAPERS
DESIGN OF STEEL AND R.C.C STRUCTURE 3 30 70 100 28 40 27 A 27 F
ESTIMATING AND COSTING 3 30 70 100 28 40 28 54 82 A
WATER RESOURCES ENGINEERING. 3 30 70 100 28 40 27 46 73 B
CONSTRUCTION MANAGEMENT 2 30 70 100 28 40 27 35 62 C
SOLID WASTE MANAGEMENT 2 30 70 100 28 40 28 42 70 B
-----------------------------------------------------------------------------------------------
PRACTICAL PAPERS
ESTIMATING & COSTING LAB 2 15 35 050 - 20 14 32 46 A+
DESIGN OF STEEL AND RCC STRUCTURE 2 20 30 050 - 20 18 27 45 A+
LAB
-----------------------------------------------------------------------------------------------
TERM WORK PAPERS
IN-PLANT TRAINING 2 15 35 050 - 20 14 31 45 A+
MINOR PROJECT 2 07 18 025 - 10 07 16 23 A+
COURSE UNDER MOOCS /NPTEL / OTHERS 1 20 30 050 - 20 18 28 46 A+
COURSE PRIMAVERA/ 3D MAX / OTHERS 1 07 18 025 - 10 06 17 23 A+
-----------------------------------------------------------------------------------------------
GRAND TOTAL 23 750 542
-----------------------------------------------------------------------------------------------
SGPA
-----------------------------------------------------------------------------------------------
REMARKS : Promoted with Carry in - DESIGN OF STEEL AND R.C.C STRUCTURE

Result Published On: 18/05/2023


A Voting Approach Applied to Preventive Maintenance… 545

Daher SFD, de Almeida AT (2012) The use of ranking Veto concept to mitigate the compensatory effects
of additive aggregation in group decisions on a water utility automation investment. Gr Decis Negot
21(2):185–204. doi:10.1007/s10726-011-9266-2
d’Angelo A, Szidarovsky F, Eskandary A (1998) Social choice procedures in water resource management.
J Environ Manag 52(3):203. doi:10.1006/jema.1997.0156
de Almeida AT, Cavalcante CAV, Alencar MH, Ferreira RJP, Almeida-Filho AT, Garcez TV (2015b)
Multicriteria and multi-objective models for risk, reliability and maintenance decision analysis. In:
International Series in Operations Research & Management Science. Springer, New York
de Almeida AT (2012) Multicriteria model for selection of preventive maintenance intervals. Qual Reliab
Eng Int 28:585–593. doi:10.1002/qre.1415
de Almeida AT, Ferreira RJP, Cavalcante CAV (2015a) A review of the use of multicriteria and multi-
objective models in maintenance and reliability. IMA J Manag Math 26(3):249–271. doi:10.1093/
imaman/dpv010
de Almeida AT, de Almeida JA, Costa APCS, de Almeida-Filho AT (2016) A new method for elicitation
of criteria weights in additive models: flexible and interactive tradeoff. Eur J Oper Res 250:179–191.
doi:10.1016/j.ejor.2015.08.058
Draginčič J, Vranešević M (2014) AHP-based group decision making approach to supplier selection of
irrigation equipment. Water Resour 41:782–791. doi:10.1134/S0097807814060050
Glasser GJ (1969) Planned replacement: some theory and its application. J Qual Technol 1(2):110–119
Halabi AX, Montoya-Torres JR, Obregón N (2012) A case study of group decision method for environmental
foresight and water resources planning using a fuzzy approach. Gr Decis Negot 21(2):205–232. doi:10.
1007/s10726-011-9269-z
Hämäläinen RP, Kettunen E, Ehtamo H, Marttunen M (2001) Evaluating a framework for multi-stakeholder
decision support in water resources management. Gr Decis Negot 10:331–353. doi:10.1023/A:
1011207207809
Hersch PL, Pelkowski JE (2014) Voter demand for fluoridated water: a tale of two c(av)ities. Appl Econ
Lett 21:51–54. doi:10.1080/13504851.2013.837573
IEEE (2012) Std 1366–2012: IEEE guide for electric power distribution reliability indices. IEEE, New York
Jiang R, Ji P (2002) Age replacement policy: a multi-attribute value model. Reliab Eng Syst Saf 76(3):311–
318. doi:10.1016/S0951-8320(02)00021-2
Keeney RL, Raiffa H (1976) Decisions with multiple objectives: preferences and value tradeoffs. Wiley,
New York
Latinopoulos D (2014) Using a choice experiment to estimate the social benefits from improved water
supply services. J Integr Environ 11:187–204. doi:10.1080/1943815X.2014.942746
Li F, Brown RE, Freeman LAA (2003) A Linear Contribution factor model of distribution reliability
indices and its applications in Monte Carlo simulation and sensitivity analysis. IEEE Trans Power
Syst 18:1213–1215. doi:10.1109/TPWRS.2003.814906
Lienert J, Scholten L, Egger C, Maurer M (2014) Structured decision-making for sustainable water
infrastructure planning and four future scenarios. EURO J Decis Process 3:107–140. doi:10.1007/
s40070-014-0030-0
Madani K, Read L, Shalikarian L (2014) Voting under uncertainty: a stochastic framework for
analyzing group decision making problems. Water Resour Manag 28:1839–1856. doi:10.1007/
s11269-014-0556-8
Mann KB, Berry KA, Bassett S, Chandra S (2013) Voting on floodplain conservation: the role of public
values and interactions along the Carson River, Nevada. Soc Nat Resour 26:568–585. doi:10.1080/
08941920.2012.713449
Monte MBS, de Almeida-Filho AT (2016) A multicriteria approach using MAUT to assist the maintenance
of a water supply system located in a low-income community. Water Resour Manag. doi:10.1007/
s11269-016-1333-7
Morais DC, de Almeida AT, Figueira JR (2014) A sorting model for group decision making: a case study
of water losses in Brazil. Gr Decis Negot 23:937–960. doi:10.1007/s10726-012-9321-7
Morais DC, de Almeida AT (2007) Group decision-making for leakage management strategy of water
network. Resour Conserv Recycl 52:441–459. doi:10.1016/j.resconrec.2007.06.008
Morais DC, de Almeida AT (2010) Water network rehabilitation: a group decision-making approach. Water
SA 36:487–493. doi:10.4314/wsa.v36i4.58425
Morais DC, de Almeida AT (2012) Group decision making on water resources based on analysis of individual
rankings. Omega 40:42–52. doi:10.1016/j.omega.2011.03.005

123
546 A. T. de Almeida-Filho et al.

Nurmi H (1983) Voting procedures: a summary analysis. Br J Polit Sci 13(2):181–208. doi:10.1017/
S0007123400003215
Nurmi H (2014) Some remarks on the concept of proportionality. Ann Oper Res 2015:231–244. doi:10.
1007/s10479-012-1252-9
Ostrom E (1990) Governing the commons: the evolution of institutions for collective action. Cambridge
University Press, Cambridge
Rausand M, Høyland A (2004) System reliability theory: models and statistical methods. Wiley, New Jersey
Recife (2016) Characterization of neighborhoods of recife. Demographic census of 2010 (in Portuguese).
http://www2.recife.pe.gov.br/servico/perfil-dos-bairros. Accessed 27 May 2016
Roozbahani A, Zahraie B, Tabesh M (2012) PROMETHEE with precedence order in the criteria (PPOC)
as a new group decision making aid: an application in urban water supply management. Water Resour
Manag 26(12):3581–3599. doi:10.1007/s11269-012-0091-4
Schoolmaster FA (1984) A geographical analysis of voting patterns for water related constitutional amend-
ments in Texas, 1957–1981. J Am Water Resour Assoc 20:151–162. doi:10.1111/j.1752-1688.1984.
tb04665.x
Silva VBS, Morais DC, de Almeida AT (2010) A multicriteria group decision model to support watershed
committees in Brazil. Water Resour Manag 24:4075–4091. doi:10.1007/s11269-010-9648-2
Smith JH (1973) Aggregation of preferences with variable electorate. Econometrica 41:1027–1041. doi:10.
2307/1914033
Srdjevic B et al (2007) Linking analytic hierarchy process and social choice methods to support group
decision-making in water management. Decis Support Syst 42:2261–2273. doi:10.1016/j.dss.2006.
08.001
Srdjevic Z, Srdjevic B (2013) Introducing the social participation concept in water management in Serbia,
and related decision-making issues. Water Resour 40:469–475. doi:10.1134/S009780781304012X
Straffin PD (1980) Topics in theory of voting. Birkhäuser, Cambridge
Sun B, Wei M, Du J, Ji W, Wen M (2015) Multi-attribute group decision making method of ecological water
compensation program based on preference of decision makers. J Coast Res 73:606–610. doi:10.2112/
SI73-105.1
Trojan F, Morais DC (2012) Prioritizing alternatives for maintenance of water distribution networks: a group
decision approach. Water SA 38:555–564. doi:10.4314/wsa.v38i4.11
UN – United Nations (2016) Resolution 64/292: the human right to water and sanitation. http://www.un.
org/ga/search/view_doc.asp?symbol=A/RES/64/292. Accessed 27 May 2016
Villaverde AR, Tadeo AJP, Gómez FG (2015) The ’social choice’ of privatising urban water services: a case
study of Madrid in Spain. J Policy Model 37:616–629. doi:10.1016/j.jpolmod.2015.04.005
Weibull W (1951) A statistical distribution function of wide applicability. ASME J Appl Mech 293–297
Zhao Y, Zhang F, Zhang D (2013) A group decision making model of water resources management based on
supporting degrees of experts. Appl Mech Mater 357–360:2362–2365. doi:10.4028/www.scientific.
net/AMM.357-360.2362

123

You might also like