Abstract: Aiming at the minimum materials handing costs and the maximum utilization rate of occupied space, a multi-objective optimization model of facility layout was established. Since the traditional multi-objective algorithms could not solve the model, a differential cellular multi-objective genetic algorithm was proposed. Through introducing differential evolution strategy into canonical cellular multi-objective genetic algorithm, the algorithm integrated features of cellular algorithm's good diversity, differential evolution strategy's good convergence and wide covering area. To evaluate the algorithm, a comparison with the canonical cellular multi-objective genetic algorithm and NSGAII on benchmarks and workshop facility layout models was conducted. The data and performance comparison indicated that the algorithm had better convergence, diversity and expansibility for multi-constraints, multivariable and non-linear models. It could solve the the relevant practical problems effectively.

Key words: cellular topological structure, differential evolution strategy, multi-objective genetic algorithms, workshop facility layout, optimization design

摘要： 以物料搬运费用最小和车间设备占地面积利用率最大为目标，建立了车间设备布局多目标优化设计模型。针对常用多目标算法不能很好求解该模型的问题，提出一种差分元胞多目标遗传算法。该算法在经典元胞多目标遗传算法的基础上引入差分演化策略，从而集成了元胞算法多样性好和差分演化策略在解决复杂问题时收敛性强、覆盖范围广的特点。分别运用该算法、经典元胞多目标遗传算法和NSGAII对测试函数及车间设备布局模型进行计算，通过数据和性能比较分析表明，针对多约束、多变量、非线性的模型，新算法具有良好的收敛性、分布性和扩展性，能有效解决相关生产实践问题。

关键词: 元胞拓扑结构, 差分演化策略, 多目标遗传算法, 车间设备布局, 优化设计