Abstract: Considering the influence of demand uncertainty in just in time production environment,a novel analysis method for Kanban number determination based on production simulation and optimization approach was presented.The problem for optimizing Kanban number to minimize its operation cost and tardiness cost was first described.By combining the production simulation method and improved Particle Swarm Optimization (PSO) algorithm with random inertia weight,the solution model for this problem was designed.The manufacturing process simulation model was constructed to evaluate the performance of Kanban number adjustment properly.The improved PSO approach could be used to enhance the global searching ability.A case study for optimizing Kanban number of each production process of a product in a shop floor was given,and the effectiveness of the proposed method was verified.The proposed method had overcame the defects of traditional analysis methods,which neglected the demand uncertainty and made it difficult to obtain satisfied system performance.In the proposed method,the overall balance of a kanban system between its performance and operation cost was achieved,it can be used to provided supporting for Kanban number determination more properly.

Key words: just in time, Kanban number, production simulation, particle swarm optimization

摘要： 考虑准时化生产环境下外部需求的不确定性,提出一种看板数量决策的生产仿真优化分析求解方法。以看板运行成本与延期成本之和最小为目标,建立了看板数量优化问题模型。结合生产仿真模型与改进粒子群优化算法(引入随机策略更新惯性权重),构建了问题求解模型。生产仿真模型用于对看板数量调整效果做出准确评价,通过改进粒子群算法控制求解流程,可获得较强的全局寻优能力。以某产品生产线为例,对各工序看板数量进行了优化求解,验证了模型方法的有效性。该方法克服了传统经验分析法忽视需求不确定性、看板调整效果难以保证的缺陷,实现了看板系统整体性能与运行成本的平衡,能够为看板数量精准优化决策提供支持。

关键词: 准时生产, 看板数量, 生产仿真, 粒子群优化