Abstract: To balance the contradictory between product diversity and manufacturing complexity of parallel mixed-model assembly system,the system complexity evaluation method was proposed based on information entropy,and the product process of customers'choice was illustrated according to the customer utility.Multi-objective optimization model which combined with the requirement of system balance efficiency was constructed by considering two objectives such as maximizing the market share of product family and minimizing manufacturing complexity.A two layer genetic algorithm for solving this kind of problem was proposed.The optimal product mix and scheduling scheme were acquired under the given objective.The feasibility and effectiveness of proposed approach was verified by the contrast of numerical examples.

Key words: customer preference, product variety, manufacturing complexity, parallel mixed-model assembly, scheduling, two layer genetic algorithm

摘要： 为平衡产品多样性和制造系统复杂性之间的矛盾,以并行装配系统为研究对象,提出基于信息熵的系统复杂性评价方法,并根据顾客需求效用描述顾客选择产品的过程,结合并行系统平衡效率的要求,构建以产品族所占市场份额最大及制造复杂性最小为目标的多目标优化模型;提出针对该问题的双层遗传算法,获得同时满足高市场占有率和低制造复杂性要求的最优产品组合和优化调度方案。通过算例的对比分析验证了所提方法的可行性和有效性。

关键词: 顾客偏好, 产品变体, 制造复杂性, 并行混流装配, 调度, 双层遗传算法