On the Improved Genetic Algorithm Based on the Strategy Pool and Expansion Mechanism for Solving Traveling Salesman Problem
Addressing the issues of low optimization efficiency and premature convergence encountered by the traditional ge-netic algorithm(GA)in solving the Traveling Salesman Problem(TSP),which are caused by the loss of population diversity and weakened local search capabilities,an improved genetic algorithm based on the strategy pool and expansion mechanism(SPEM-IGA)is proposed.Two sets of strategy pools are designed for different purposes.A local search strategy pool,consisting of 2-opt,heuristic insertion,and greedy operator,is constructed to enhance the search depth of solutions.To broaden the search scope of solutions,a global search strategy pool is formed by combining operators such as the nearest neighbor insertion,flip,seg-ment exchange,and circular left shift.Considering the level of population diversity,a random selection mechanism is designed based on the strategy pool,which dynamically expands the population,effectively improves its diversity and balances its capability for both global and local search.Superior individuals in the population are retained through elite selection to accelerate the conver-gence speed of the algorithm.Simulation experiments show that the improved genetic algorithm based on the strategy pool and ex-pansion mechanism has better solution accuracy and stability compared to the existing literature.
国家科技期刊平台
NSTL
万方数据
