一种改进的变种群规模遗传算法

传统的遗传算法在遗传操作的过程中,不管具体问题对种群规模的要求都事先指定,造成种群规模表现的不合理。针对这一问题,Arabas,J.等人提出了一种变种群规模遗传算法GAVaPS(Genetic Algorithms withVarying Population Size),旨在遗传操作的过程中自适应地调整种群的规模。黄乐等人提出了一种改进的变种群规模遗传算法VPSG-GM(Varying Population Size Genetic Combined with Gradient Search Method),它与GAVaPS方法在遗传操作上的不同之处是把评价目标函数的步骤置于重组种群时“淘汰”步骤之前,这样做减少了算法由于重组造成的不必要开销。文章提出的改进的变种群规模的遗传算法,在调整种群规模时通过增加补算算子,从而达到保持种群多样性的目的。这样做可以略去变异操作,减少了系统开销。实验证明了该算法的有效性。

An Improved Genetic Algorithm with Varying Population Size

The traditional genetic algorithm appoints the population size,regardless of actual question.So it causes the unreasonable population size.According to this,Arabas,J.proposed GAVaPS(Genetic Alogorithms with Varying Population Size) which aims at adjusting population size self-adaptively.Huang Le proposed VPSG-GM(Varying Population Size Genetic Combined with Gradient Search Method).The difference between GAVaPS and VPSG-GM is to put the step of evaluating function before assembling population.It can reduce unnecessary costs used to assemble population.This paper proposes an improved genetic algorithm with varying population size that achieved the goal of keeping the variety of the population through adding complementary operator in adjusting population size.What we do can get rid of mutating operation and reduce the system cost.The experimental results show that the algorithm we propose is effective.