基于遗传算法的单件准时化生产计划

设计了一种遗传算法，用来解决以准时生产为目标的单件制造业生产计划问题，提出了基于投产日期的生产计划编码方法，这种编码方法直接把生产计划映射为一个有序的字符串，并针对这种编码设计了专门的交叉算子和变异算子，相似度检验增加了初始种群的多样性；小生境进化技术的运用有效地保持了群体的分布性；种群并行进化策略、变异率自适应调整、与启发式算法相结合等措施提高了遗传算法收敛到最优解的成功率。

An Effective Genetic Algorithm for Just-in-time Production Planning in Job Manufacturing Systems

A simple and universal gene encoding scheme for just-in-time production planning in job manufacturing systems was presented, by which a production planning directly corresponds to an ordering string. For the encoding, three important genetic operators: selection, crossover and mutation were contrived and described in derail. In order to increase the algorithm's capacity of global convergence, some strategies such as adaptive probability of mutation, calculating the degree of similarity, niche technology based on preselection, multigroup parallel and combining with heuristic approach were adopted.