项目名称: 基于混合差分进化的多目标工艺规划和调度研究
项目编号: No.U1304609
项目类型: 联合基金项目
立项/批准年度: 2014
项目学科: 自动化技术、计算机技术
项目作者: 张闻强
作者单位: 河南工业大学
项目金额: 30万元
中文摘要: 工艺规划和调度以其基础性、重要性以及复杂性,成为柔性制造系统领域研究的重点和热点。而多目标优化因其复杂性、现实性以及给决策者提供更多更实际的备选方案集等特征,成为工艺规划和调度研究中的难点。研究此类实际生产调度中典型组合优化问题的多目标进化算法具有重要的学术意义和应用价值。本项目围绕工艺规划和调度的多目标优化,通过创新的混合差分进化框架构建、特殊的适应度评价函数设计、新颖的混合选择机制、独特的精英保存策略以及基于多个目标的混合局部搜索技术等,以增强算法在Pareto前沿面多个方向上的收敛性能并保证算法的均匀分布性能,同时尽可能的降低算法的时间复杂度。本研究旨在混合进化算法框架设计、方向性搜索、收敛和分布性能平衡以及效率提高等方面取得一系列有指导价值的理论与算法应用成果。本项目对于混合差分进化解决多资源约束条件下的多目标工艺规划和调度问题的研究,将有利于推动复杂智能加工制造系统的研究和发展。
中文关键词: 工艺规划和调度;差分进化;多目标优化;收敛和分布;
英文摘要: Process planning and scheduling (PPS) is an important and hot research topic in flexible manufacturing system (FMS) due to its complexity, fundamentality and importance. Owing to the complicated and practical characteristics as well as the ability of providing the many Pareto optimal solutions with incommensurable quality for decision makers, multiobjective optimization is great difficulty problem in PPS. Therefore, it will provide significant impacts on academic and application areas to do research on multiobjective evolutionary algorithm (MOEA) for such typical combinatorial optimization problem in real-world production scheduling. This project major concerns the multiobjective optimization of PPS and attempts to propose a new hybrid differential evolution (DE) framework, a special design of fitness evaluation function, a novel hybrid selection mechanism, a distinct elitism preservation strategy and a masterly combined local search technology basing on multiple objectives, to enhance the convergence and distribution performances, while reducing the time complexity of the algorithm as much as possible. The main achievements of this study lie in the series of guidance values at theory and application of algorithms for designing of hybrid MOEA framework, searching with multiple directions, balancing of convergenc
英文关键词: Process Planning and Scheduling;Differential Evolution;Multiobjective Optimization;Convergence and Distribution;