时空耦合约束下随机装配线再平衡方法研究

The built assembly line should be re-planned and rebuilt as a result of the changes of market demand, product features and manufacturing process. The assembly line rebalancing problem is one of the important cores of the assembly line re-planning. This project aims at reducing the cost of the assembly line modifications, improving the efficiency and stability of the modified assembly line. The model of the stochastic assembly line rebalancing problem with the time and space coupling constraints is constructed, including the probability distribution of the task processing time and the space constraints of the previous assembly line. By using the time and the space properties of the tasks, the distance function in the feasible solution space of the assembly line rebalancing problem is defined to describe the similarity of the different solutions. The hybrid Pareto imperialist competitive algorithm is proposed to find the Pareto optimal solutions set for the assembly line rebalancing problem with the multi-objectives, by combining the Pareto dominant strategy and the principle of the immune algorithm. The simulation based on the assembly line meta-model is implemented to analyze the performance of the assembly lines computed by the hybrid imperialist competitive algorithm, and the interactive feedback strategy is proposed to adjust the configuration of the workers, the buffers and the parallel stations. Then the evolved assembly line rebalancing model and the corresponding evolving feedback strategy are both developed to reassign the tasks among the stations. Based on the simulation and analysis of the assembly lines, the interactive feedback strategy and the evolving feedback strategy are combined to improve the assembly line. Finally, the prototype system for the assembly line rebalancing problem is developed to test the proposed theories and methods, which can support to rebuild the efficient and stable assembly lines.

市场需求、产品和工艺等变化环境下，原先建造的装配线需重新规划和改建。装配线再平衡是装配线重新规划的重要核心内容之一。本项目以降低装配线的改建成本，提高装配线效率和稳定性为目标，建立装配任务时间的随机变化规律和原有装配线的空间模型，研究时空耦合约束下多目标随机装配线再平衡模型。利用装配任务的时间和空间属性，定义解空间的距离函数，研究再平衡的解空间分布规律。将Pareto支配关系和免疫算法原理融入帝国竞争算法，设计混合Pareto帝国竞争算法。研究基于装配线元模型的快速建模方法，提出基于装配线仿真的交互式反馈策略，调整缓冲和工位等设施布置；然后，构建装配线再平衡的进化模型，研究基于装配线仿真的进化式反馈策略，再次优化任务分配，形成交互式反馈和进化式反馈的循环机制，实现装配线再平衡的全面优化。开发装配线再平衡原型系统，验证所提出的理论模型和方法，为大量的装配线升级改造提供理论指导和实践工具。

市场需求、产品和工艺等变化环境下，原先建造的装配线需重新规划和改建。装配线再平衡是装配线重新规划的重要核心内容之一。本项目综合考虑装配线的时间和空间约束，构建了时空耦合约束下的双边装配线再平衡数学模型。针对双边装配线负荷均衡的问题，提出了工位完工时间偏差的评价指标，设计了一种基于Moodie and Young的改进启发式算法。针对装配线再平衡过程中涉及到的工人再分配问题，考虑工人技能等级对任务作业时间的影响，提出了一种启发式规则算法，确保工人的再分配不对装配线的效率产生负效果。针对需求变化下的双边装配线再平衡问题，构建了考虑优化效率与成本两个目标函数的数学模型，并设计了一种基于任务移动的启发式规则指导任务的再分配。针对工艺变化下的双边装配线再平衡问题同样构建了以优化效率与成本为目标的数学模型，并设计了一种改进的NSGA-II算法。其采用变异与规则相结合的初始种群生成方法、基于单工位任务分配规则的解码方法、基于改进的最佳匹配规则的交叉变异方法和比例选择与锦标赛选择相结合的选择规则，为了增加种群多样性，还设计了相似个体交叉限制以及去重方法。针对时空约束下的双边装配线再平衡问题，提出了一种改进的帝国竞争算法，改进了帝国初始化、解码、同化和帝国竞争的机制，改善其性能。之后，考虑工程实际上任务作业时间随机波动等问题，提出了基于仿真反馈策略的装配线再平衡方法。最后，在以上理论研究的基础上，开发了装配线再平衡原型系统。应用于某公司工程机械装配线和三家公司的发动机装配线改建，验证了项目研究中提出的模型与方法的有效性。