基于改进量子进化算法的不常用备件联合采购与配送调度协同优化研究

Rarely-used spare parts are important for the normal maintenance and emergency treatment of equipments. The collaborative optimization strategy of Joint Replenishment and Delivery Scheduling(JRD) is an effective way of cost control for the continuous production enterprises that are in front of global purchasing environment in China.The key success factors of this strategy are the rationality of the treatment of uncertain factors and the scientificalness of the design of collaborative strategy. Most of the existing researches are deterministic model (the assumption is not practical) and stochastic model (no necessary sample to test the assumption). Moreover, sharing of spare parts within a speific industry is not considered because of the complexity of algorithm. In this project, we will design novel membership fuctions to characterize the variables firstly.Secondly, two JRD collaborative optimization models will be designed considering industry sharing, delivery grouping constraints,and practical resource constraints. Thirdly, we will analyze the properties of optimal strategy and provide novel approaches with good robustness, highly accuracy, and fast convergence. These approaches use hybrid adaptive quantum evolution algorithms which integrate the advantages of differential evolution/particle swarm optimization algorithm and design skills of existing heuristic algorithm. At last,the proposed models and algorithms will be tested by the case studies in nuclear power plants and hydropower stations.This project is a crossed research of novel intelligent algorithms and JRD problem with important theory significance and high practical value.

不常用备件是设备维护和应急处理的重要保障，对面向国际采购环境的我国连续性生产企业而言，对联合采购与配送调度(JRD)进行协同优化可有效地控制成本，其关键在于合理处理复杂的不确定因素和设计科学的协同优化策略。现有研究多是确定性(过于理想)和随机性优化模型(常缺乏必须的样本来客观地验证假设是否成立)，且没有考虑到行业共享，原因之一在于求解非常复杂。本项目将设计新颖的模糊隶属度函数来刻画不确定变量；考虑行业共享、配送分组约束和贴近管理实践的资源约束，构建两类JRD协同优化模型；分析模型最优策略性质，设计有效融合自适应量子进化算法(吸收差分进化/粒子群优化算法的优点)和启发式算法的智能求解新方法(稳健性好、精度高、收敛快)；结合核电/水电企业进行应用研究，验证模型和算法的科学适用性。项目属于新颖的智能算法与联合采购-配送问题的交叉研究，理论意义重要；将解决有很强现实意义的协同优化难题，实用价值高。

不常用备件是设备维护和应急处理的重要保障，对面向国际采购环境的我国连续性生产企业而言，对联合采购与配送调度(JRD)进行协同优化可有效地控制成本，其关键在于合理处理复杂的不确定因素和设计科学的协同优化策略。本项目该出刻画不确定变量的方法；提供了基于量子进化、果蝇优化算法、定界和变领域搜索相结合的基本JRD高效求解算法(稳健性好、精度高、收敛快)；考虑数量折扣和资源约束、多仓库、配送过程优化、与选址相结合等管理实情，构建四种改进的JRD协同优化模型并提供有效求解算法；结合核电/水电企业进行应用研究，验证了模型和算法的科学适用性。项目属于新颖的智能算法与联合采购-配送问题的交叉研究，理论意义重要；将解决有很强现实意义的协同优化难题，实用价值高。目前发表和录用SCI和SSCI期刊论文16篇、EI期刊论文1篇，依托项目毕业博士生4人和硕士生6人。