Genetic Algorithm Based Multi-objective Optimization Framework to Solve Traveling Salesman Problem

Abstract

The combination of objectives to be optimized is known as Combinatorial Optimization. Traveling Salesman is one of the demanding topics in the field of Combinatorial Optimization whereas in traditional TSP, only unique objectives are optimized. In this research work, the Multi-objective Traveling Salesman Problem is considered. The multi-objective TSP is an expanded instance of TSP. In MOTSP, more than one objective function is determined and optimized to meet the finest solutions. MOTSP is solved by using Genetic Algorithm. MOTSP gave the best possible Pareto optimal solutions for all the data instances that were tested. Genetic Algorithm produces near-optimal solutions in a fair period of time for almost all the objectives. The Traveling Salesman Problem was tested with various instances from TSPLIB dataset which consists of different number of cities in this work. The experimental results showed that the GA gives nearest optimal solutions to the problem instances taken from the TSP Library and considered in this work.