提高变型标准遗传算法收敛速度的混合法及其推广

HYBRID APPROACH FOR RAISING CONVERGENCE RATE OF VARIANTS OF CANONICAL GENETIC ALGORITHMS AND ITS GENERALIZATION

摘要: 针对即使是全局收敛的变型标准遗传算法VCGA¹有时也会发生收敛速度变慢的问题，提出了VCGA和最陡下降法相结合的混合法HVCSDA．该方法增强了VCGA在接近全局最优解时的微调能力．还将HVCSDA推广到一类修正的VCGA上．仿真实例表明了HVCSDA及其推广能有效地提高收敛速度．对30城市TSP的仿真结果为6.822，要好于用TABU得到的6.99的结果⁶．

Abstract: Aimed at the problem that VCGA with global conver gence so metimes slows the conver gence rate down,this paper suggests a hybrid approach called HVCSDA which can reinforces the capability of fine tunning of VCGA as it is near the global optimal solution. This paper also gener alizes HVCSDA in order to adapt for aclass of the modified VCGA. Two simulation examples show that HVCSDA and its gener alization can effectively raize the converg encerate of (aclass of modified)VCGA. In the second simulation example,the length of routing of 30 cities TSP is 6.882,it is better than one that is 6.99 by TABU.