基于遗传蝙蝠算法的选择性拆卸序列规划

针对产品选择性拆卸序列规划问题，提出一种基于遗传蝙蝠算法的产品拆卸序列规划方法.利用Python语言对传统蝙蝠算法进行离散化处理，并在种群更新过程中引入遗传算法的交叉与变异机制，生成遗传蝙蝠算法，以增强解搜索的多样性；在构建适应度函数模型时以拆卸工具的变化次数与拆卸方向的重新定位次数作为评价指标，同时加入零部件的回收收益指标，使适应度函数更加完善.以工业机械臂为实例，利用所提方法进行产品拆卸序列规划求解，对比传统蝙蝠算法以及遗传算法的求解结果，发现在一定的种群数目下，所提方法收敛时间较短；在不同种群数目下，所提方法得到的适应度函数最优值质量较高，从而验证了遗传蝙蝠算法的搜索优越性.
Abstract: A method based on genetic-bat algorithm (GBA) was proposed to resolve the selective disassembly sequence planning (SDSP) problem. The traditional bat algorithm was discretized by Python and the crossover mutation mechanism of genetic algorithm was introduced in the process of population regeneration to generate GBA and to improve the diversity of the search algorithm. The fitness function model was constructed, with the recovery benefit of the disassembly parts, changes of disassembly directions and disassembly tools as the evaluation indexes. The industrial mechanical arm was studied as an instance to compare traditional bat algorithm (BA) and genetic algorithm (GA). Results showed that the convergence time of GBA was shorter when the population number was defined, and under different population numbers, the optimal value of fitness function of GBA was higher. The superiority of the GBA was verified.