基于浮点型编码策略的差分多目标柔性车间调度优化

运用进化算法求解柔性车间调度问题时,编码的特殊性对进化策略造成的局限制约了算法的搜索能力。为此,提出一种基于浮点型编码策略的差分多目标优化算法。该算法采用基于工序权重的浮点数编码—解码机制,消除了排列组合型编码方式对进化操作带来的约束,运用差分进化策略生成新个体,以提高优秀个体产生的几率,进而保证算法有更好的收敛性。将算法与传统算法及其改进形式在相同测试用例上进行对比,结果表明,本算法在保证收敛性的同时,搜索到更多的非支配个体,体现出更好的分布性。此外,提出了平行决策和等价平行决策的定义,将柔性车间调度模型的研究拓展至决策空间。

Float-code based differential evolutionary multi-objective optimization forflexible Job-Shop scheduling problem

Evolutionary algorithms (EAs) for optimizing flexible Job-Shop scheduling are always confined to the code-decode mechanism. Traditional coding strategy restricts the search ability of EAs. This paper proposed a differential evolutionary algorithm using float value based coding strategy on operation procedures for solving FJSP. Combining the float-based weight coefficient code-decode mechanism and differential evolutionary strategy, the algorithm averted the influence of permutation and combination coding strategies and could obtain the global optimal solutions. The results obtained from the computational study show that this algorithm can achieve the Pareto optimal solution with an ideal uniformity. Furthermore, giving the definitions of parallel decision and equivalent parallel decision expands the study of FJSP into decision space.