提出了一种自调节种群的演化算法(SaPEA)求解旅行商问题,算法根据当前最优适应度改进的情况提出一种更精细调节种群规模的模式,并根据演化的进程选择强化操作或者分化操作。这样不仅有利于保持种群的多样性开发新的解,还可以加快收敛速度探索到更好的解。同时,还对现有的启发式交叉算子和3-opt局部搜索算法进行了改进。通过对TSPLIB中实例进行测试,表明了SaPEA算法的优越性。
A Self-adjusting Population Evolutionary Algorithm (SaPEA) for traveling salesman problem was proposed. The algorithm develops a refined scheme for adjusting population size based on improvement of best-so-far fitness, which also combines with intensification or diversification based on evolutionary process. SaPEA not only contributes to maintain the diversity of the population for exploring new solutions, but also speedups convergence for exploiting better solution. Meanwhile, improvements on heuristic crossover and 3-opt local search were implemented. Tests on examples of TSPLIB indicate the efficiency of SaPEA.