Abstract
The purpose of this study to analyze genetic algorithm (GA) and simulated an-nealing (SA) based approaches applied to well-known Traveling Salesman Prob-lem (TSP). As a NP-Hard problem, the goal of TSP is to find the shortest route possible to travel all the cities, given a set of cities and distances between cities. In order to solve the problem and achieve the optimal solution, all permutations need to be checked, which gets exponentially large as more cities are added. Our aim in this study is to provide comprehensive analysis of TSP solutions based on two methods, GA and SA, in order to find a near optimal solution for TSP. The re-sults of the simulations show that although the SA executed with faster comple-tion times comparing to GA, it took more iterations to find a solution. Additional-ly, GA solutions are significantly more accurate than SA solutions, where GA found a solution in relatively less iterations. The original contribution of this study is that GA based solution as well as SA based solution are developed to perform comprehensive parameter analysis. Further, a quantifiable comparison is provided for the results from each parameter analysis of GA and SA in terms of performance of solving TSP.
Highlights
In recent years, the importance of Artificial intelligence (AI) keeps increasing, and there is a new application of AI arises every day
It would be prudent to define an aspect of this test that is shared among genetic algorithm (GA) and simulated annealing (SA) for simulation experiments
Since these algorithms have been adapted to be utilized by Traveling Salesman Problem (TSP), some terms that have been adapted have more to do with TSP than the actual algorithm being leveraged upon
Summary
The importance of Artificial intelligence (AI) keeps increasing, and there is a new application of AI arises every day. The GA and SA are used in wide range of application areas, such as Bayesian inference links to particle methods in Bayesian statistics and hidden Markov chain models, electronic circuit design, learning fuzzy rule base, to train neural networks, and more. The exact algorithms are guaranteed to find the optimal solution, in which number of iterations needed increases exponentially. There are several studies in the literature on heuristic algorithm-based TSP solution, some of them worth mentioning here. The approach presented contains most of the heuristic algorithms used to test TSP, including greedy algorithm, minimum-cost spanning tree, and local search. Modern optimization methods were included such as SA, GA, ant algorithm, particle swarm optimization, tabu search algorithm, Hopefield neural networks [5]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
