A New Hybrid Genetic and Simulated Annealing Algorithm to Solve the Traveling Salesman Problem

The Traveling Salesman Problem (TSP) is one of the most widely known non- deterministic polynomial (NP-hard) problems. It is used to check the efficacy of any combinatorial optimization method, and is often used as a testbed. This paper proposes a new Hybrid Genetic and Simulated Annealing Algorithm (HGSAA) to solve the TSP. The basic idea behind the proposed HGSAA is that the genetic algorithm (GA) starts its process with its new techniques. When it seems that the GA is stuck after 20 consecutive generations, the hybrid algorithm switches the population to the simulated annealing (SA) which allows uphill jumps to a higher-cost solution in order to avoid getting trapped in local minima. After several SA iterations; the new population is returned to GA. The proposed algorithm was tested using benchmark datasets for symmetric TSPs from TSPLIB, and it provided good results within a reasonable time.