Modeling TSP with Particle Swarm Optimization and Genetic Algorithm
Shaukat Ali Khan, Sohail Asghar, Simon James Fong · Advanced Information Management and Service · 2010
Traveling Salesman Problem (TSP) is a classical problem of optimization for researchers and its modeling is of great interest for Engineering, Operations Research and Computer Science. For solving TSP, many methods have been proposed, including heuristic ones. Our work extends the hybrid model, based on Particle Swarm Optimization, Genetic Algorithms and Fast Local Search, for the symmetric blind travelling salesman problem proposed by Thiago R. Machado and Heitor S. Lopes. We have replaced the fast local search with mutation to avoid the overhead for finding optimal solution. We have also replaced the one point crossover with uniform crossover as one point crossover is found to generate invalid tours in most of the cases. We argue that this model is more efficient as compared to the model proposed by Thiago R. Machado and Heitor S. We implement a prototype of the model and show its feasibility.