传统遗传算法中染色体的编码形式一般为链条形,即不论采用二进制编码还是实数编码,可行解均以链条的形式表现,遗传操作也是在这种链式编码的基础上进行的。当决策变量增多,链条加长时,这种遗传算法的计算效率变得很低。此外,在梯级水库优化调度中,由于上、下游水库间存在的耦合关系,使得上游水库基因段中某一位基因的改变将连锁式地引发下游各水库基因段中相应基因的改变,这种连锁变化在链式编码中的实现是较复杂的。为此,本文提出了矩形体编码的遗传算法,它可以有效提高传统遗传算法在处理这类问题时的效率,并使优化结果得到提高。最后通过一个算例,说明了该方法的可行性与有效性。
In traditional genetic algorithm(GA),chromosomes coding is usually in chain form that is the form of feasible solutions.This means that binary or real-number encoding may take such a chained coding and it may also be used for genetic computing.One shortcoming of the chained coding is that the length of chromosome chain will become longer when the number of decision variables increases,leading to a much lower efficiency of computation.Another is the coding difficulty caused by coupling of the upstream and downstream reservoirs,as a change in any gene of the upstream reservoir' chromosomes will result in consecutive changes in the downstream chromosomes.This paper presents an advanced GA of rectangular coding that by resolving these problems improves greatly the computation efficiency and the optimization results.A simulation example is given to verify the feasibility and validity of the proposed algorithm.