基于遗传算法的自动装配顺序规划

笔者对自动装配顺序规划问题建立一种改进的遗传算法（Genetic Algorithms)。建立了装配体各零件间的可行装配方向元组，用于判断无干涉装配序列；提出5种合适的遗传算子，并以装配方向的改变次数作为装配顺序的适应值；采用一种适应值修改方案来避免算法过早收敛。算例表明，本文建立的装配顺序规划的遗传算法，能快速自动产生出满意的装配顺序方案。

Automatically Assembly Sequence Planning Based On Genetic Algorithms

A modified genetic algorithm for automatically assembly sequence planning is presented in this paper. The assembly direction sets of parts in structure are built to evaluate a assembly sequence without collisions. Five genetic operators are proposed. The fitness of sequence is calculated by its reorientation number, and a fitness scaling scheme is used to overcome premature termination. Example is provided to demonstrate the performance of this algorithm.