A Novel Approach for Combining Genetic and Simulated Annealing Algorithms

Abstract

The Traveling Salesman Problem (TSP) is the most well-known NP-hard problem and is used as a test bed to check the efficacy of any combinatorial optimization methods. There are no polynomial time algorithms known that can solve it, since all known algorithms for NP-complete problems require time that is excessive to the problem size. One feature of Artificial Intelligence (AI) concerning problems is that it does not respond to algorithmic solutions. This creates the dependence on a heuristic search as an AI problem-solving technique. There are numerous examples of these techniques such as Genetic Algorithms (GA), Evolution Strategies (ES), Simulated Annealing (SA), Ant Colony Optimization (ACO), Particle Swarm Optimizers (PSO) and others, which can be used to solve large-scale optimization problems. But some of them are time consuming, while others could not find the optimal solution. Because of this many researchers thought of combining two or more algorithms in order to improve solutions quality and reduce execution time. In this work new operations and techniques are used to improve the performance of GA (Elhaddad and Sallabi In: Lecture notes in engineering and computer science: proceedings of the world congress on engineering 2010, vol I WCE 2010, June 30–July 2, London, UK, pp 11–14, 2010), and then combine this improved GA with SA to implement a hybrid algorithm (HGSAA) to solve TSP. This hybrid algorithm was tested using known instances from TSPLIB (library of sample instances for the TSP at the internet), and the results are compared against some recent related works. The comparison clearly shows that the HGSAA is effective in terms of results and time.