针对大规模混流U型装配线平衡问题,提出一种基于多级随机分配编码的改进遗传算法。算法在编码阶段考虑操作间优先关系,采用多级随机分配编码方式,将大规模混流U型装配线问题转化成中规模或小规模混流U型装配线问题。该编码通过避免不可行随机解的产生,在编码阶段提高可行解的比率。同时,采用基于层级的单点交叉方法,在保证可行性的前提下,提高种群的多样性。算法在降低计算复杂度的同时也能准确求出问题较优解。对比实验结果表明所提出的算法能快速有效求解大规模混流U型装配线平衡问题。
Aiming at the large-size mixed-model U-type assembly line balancing problem,an improved genetic algorithm is proposed based on multilevel coding. The algorithm considers a priority in relations with the coding phase operation and transforms the large-size mixed-model U-type assembly line problems into medium-size or small-size mixed-model U-type assembly line problems with the use of multi-stage random allocation encoding. This encoding improves the ratio of feasible solutions by avoiding generating unfeasible solution. And it uses the level-based single-point crossover to improve diversity of the population in ensuring the viability. The algorithm can obtain the optimal solutions of the problem while reducing the computation complexity. Experimental comparison with existing approaches demonstrates that the proposed algorithm can solve large-size mixed-model U-type assembly line balancing problems quickly and efficiently.