Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 6

CHAPTER III

RESEARCH METHODOLOGY

In this chapter will be explained about the systematical procedure in


conducting this research. The research methodology will be used as a research
framework from the beginning of the research until the result is obtained. The
beginning of this research start with finding the topic of the research, finding
research supervisor, and pre-research discussion. The flowchart of the research will
be presented in Figure 3.1

3.1 Preparation Stage


In the preparation stage, the literature review is being done by the writer.
Literature review is done by collecting several information, theories, and
knowledge from books, journals, and online valid reference. Then the literature
review is written in the Chapter 2 of this research. The writer also conducted a field
study by visiting one of electronic appliances stores to learn the distribution strategy
of the electronic appliances store.

3.2 Model and Algorithm Development Stage


In this stage will be done the development of the model and algorithm of
the Vehicle Routing Problem With Multiple Synchronization (VRPMS) for
delivery and installation. Then the model and algorithm will be validated.

3.2.1 Model Development


The model of VRPMS is constructed in the form of mathematical
formulation and conceptual mode for delivery and installation. The constraints used
in this model to be considered to determine the route and schedule are planning
horizon in days, vehicle capacity, vehicle maximum travel distance per day,
electronic appliances sizes, electronic appliances penalty for installation lateness,
customer request location, customer time windows, customer request order and
quantity, technician location, technician maximum travel distance per day,
technician maximum request per day, and technician skills specialty. The objective
function of this model is to minimize the total cost that consist of vehicle travel
distance cost, vehicle daily cost, vehicle fix cost, technician travel distance cost,
technician daily cost, technician fix cost, and penalty cost for installation lateness.

3.2.2 Tabu Search Algorithm Development


After the mathematical model is constructed, the algorithm development
will be done using Tabu Search algorithm. The algorithm will be divided into two
stages, which are initiation stage and improvement stage.
The initial model is constructed using Nearest Neighbor (NN) to obtain the
initial feasible solution that satisfy the constraints. The calculation is divided into
two sub problem which are delivery and installation, the result from the delivery
routing will be the input to route the installation problem. Thus the route and the
schedule will be obtained for the initiation.
Then the improvement are done using Tabu Search by conducting
neighborhood search such as swap, slide, and flip for the routing and schedule.
There are three types of memories used to record the solutions which are short term
memory (tabu list), medium term memory, and long term memory. The number of
tabu list will be predetermined and it records the all current visited solutions as
many as the number of tabu list, the medium term memory will record several best
result from the tabu list, and the long term memory only records the best solution
(elite solution) found.
After the optimization using tabu search is done and the program is
established then the development of decision support system tools is created using
visual basic applications (VBA) in Microsoft Excel. The decision support system
tools is adjusted to the algorithm that has been developed. Then the model is
verified to check whether there is any error or debug occurred while running the
program. If there is an error, the error source need to be analyzed and the program
coding need to be repaired.
3.2.3 Model and Algorithm Validation
Validation of model and algorithm will be done by running the program
using data from simple problem. The output of the program then is matched with
the result of manual calculation using the same algorithm. If there is any different
between the result of the program and the manual calculation, the program coding
will be checked or reconstructed if necessary.
Then the validation is done to checked whether the model is suitable for
solving the synchronization problem with the constraints that have been stated. If
the result of the model is able to solve the problem already, then the model and
algorithm can be stated as valid and the numerical experiments can be proceed.

3.3 Numerical Experiment and Analysis Stage


After the model is already well-established, the numerical experiments and
analysis stage is done. The first basic experiments is done by using pre-constructed
data sets. Then several scenario is added such as. . . . The purpose of the running
the scenarios is to learn about the impacts of system behavior on the total cost of
the result.
The sensitivity analysis is also done to the existing model by selecting some
tabu parameters. The parameters are the number of tabu list, number of
neighborhood solution, and maximum number of iteration. The purpose of the
sensitivity analysis is to learn about the impacts of the best solution and iteration
time. The results of the experiments are then recapitulated, processed, and analyzed.

3.4 Conclusion and Suggestion Stage


In the last stage, the conclusion will answer the objectives of the research
which is stated in the initial chapter. The construction of the VRPMS and the tabu
search algorithm that can minimize the distribution cost for the electronic
appliances company will be concluded. Then, the suggestions are mode for this
research and for the future research.
START A

PREPARATION STAGE NUMERICAL EXPERIMENTS AND ANALYSIS STAGE

LITERATURE REVIEW:
Studying the concept of Running basic experiments and several scenario

(1) Electronic appliances


(2) Distribution management FIELD STUDY
Sensitivity analysis
(3) Vehicle routing problem Understanding the real process of
(4) Vehicle routing problem with delivery and installation of
multiple synchronization electronic appliances store.
constraints Analysis and interpretation
(5) Heuristics and metaheuristics
(6) Tabu search
(7) VBA implementation
(8) Research position analysis
CONCLUSION AND SUGGESTION STAGE

Conclusion and Suggestion


MODEL AND ALGORITHM DEVELOPMENT STAGE

Constructing mathematical formulation and conceptual model


for the VRPMS
END

Constructing Tabu Search algorithm based on the developed


model

Creating interface display for decision support system

No
Is the model verified?

No
Yes

Conducting validation using small data by comparing the


result of solving algorithm with manual calculation

Is the model and


algorithm valid?

Yes

Defining the data set used in the experimements

Figure 3. 1 Flowchart Methodology

One important element to be considered while purchasing or using the electronic


appliances is its power, which usually stated in watts unit. According to Cal Caly
Pomana (2004), the type of electronic appliances and its power in watts is as follow:
Figure 2. 1 Electronic appliances power
(Source: https://www.cpp.edu/~pbsiegel/sci210/appliances.pdf)

You might also like