IMPLEMENTATION OF GENETIC ALGORITHM TO COMPLETE TRANSPORTATION PROBLEMS OF INTERCITY BUS VEHICLES

Paryati, Paryati (2022) IMPLEMENTATION OF GENETIC ALGORITHM TO COMPLETE TRANSPORTATION PROBLEMS OF INTERCITY BUS VEHICLES. In: Application Artificial Intellegence and Future. Bentham Science, Singapore, pp. 1-32. ISBN 978-0357577722

Text PARYATI CHAPTERBOOK GA BAB 10.pdf Download (634kB)

Abstract

The design and implementation of a software as a tool to solve the multiple criteria transportation model with
fuzzy cost parameters has been carried out using genetic algorithms. This software is named TraFAG. The
software design uses the Waterfall methodology, which consists of analysis, design, implementation and testing.
The algorithm used is a genetic algorithm. This algorithm is based on genetic processes that exist in living
things, namely the development of generations in a natural population, gradually following the principle of
selection or who is strong who will survive. In the transportation system, the effect of congestion on
transportation means results in uncertainty in part or all of the coefficients on objective functions, such as
transportation costs or delivery times, which become uncertain. A way to deal with uncertainty in making these
decisions using fuzzy principles. The fuzzy cost parameter in TraFAG uses the Triangular Fuzzy Number (TFN).
In multiple criteria optimization, the determination of the optimal value uses the Pareto solution. The Pareto
solution is determined based on ordered fuzzy objective values. Comparison and sorting of fuzzy numbers, using
integral values. The TraFAG software is implemented in the Borland Delphi version 3.0 programming language
environment. which is the development of the Pascal language for the Window-based programming
environment. The solution of multiple criteria transportation problems can be solved by a heuristic approach
using genetic algorithms. Analysis of the results of the program shows that the processing time in the test case
will be directly proportional to the product of the number of source and destination depots with a correlation
coefficient of 0.89. The analysis also shows that the number of population is linearly proportional to each test
case to the processing time with a correlation coefficient of 0.99. The parameter  indicates the degree of
optimism will affect the result of the integral value linearly. The higher the alpha value, the greater the
transportation costs. The alpha that should be chosen is a value of 0.5, which is a moderate value so it is in a safe
condition. The alpha which yields the minimum cost for test case 2 to test case 6 is 0.1. The greater the
population, the smaller the fitness function. The greater the number of generations, the lower the transportation
costs. The results obtained are relatively stable on average in the generation 300 and above. The crossover
probability affects the fitness function. In cases 13 and 15, the crossover probability that results in the minimum
fitness function value is 0.1. The mutation probability has a lot of influence on the fitness function. In case 2 it
causes a stable mutation probability with a fitness value of 48.35.

Item Type:	Book Section
Additional Information:	Publish
Uncontrolled Keywords:	Fuzzy, Genetic Algorithm, Transportation Problems, Waterfall.
Subjects:	T Technology > T Technology (General)
Divisions:	Faculty of Engineering, Science and Mathematics > School of Engineering Sciences
Depositing User:	ST.,M.Kom PARYATI PARYATI
Date Deposited:	07 Apr 2023 09:29
Last Modified:	07 Apr 2023 09:29
URI:	http://eprints.upnyk.ac.id/id/eprint/34023

Actions (login required)

View Item