Abstract: For the Energy-Saving Job Shop Scheduling Problem (EJSP),a energy-saving scheduling model was established with the objective of optimizing the total energy-consumption and the makespan.For the model,a Multi-objective Discrete Grey Wolf Optimization (MODGWO) algorithm was proposed to solve it.According to the characteristics of the problem,a discrete integer encoding method was first employed to represent the scheduling solutions,and some dispatching rules were adopted to generate the initial population.A two-mode parallel search method was introduced based on the tracing mode and the seeking mode.The numbers of individuals in the two modes were dynamically adjusted in the search process to coordinate the global and local search capabilities of the algorithm.To make the algorithm suitable for the multi-objective discrete scheduling problem,a discrete individual updating method was presented based on crossover operation in the tracing mode,and a discrete individual updating method was developed based on memory pool mechanism and neighborhood structures in the seeking mode.40 benchmarks of the job shop scheduling problem were modified to verify the effectiveness of the proposed MODGWO.

Key words: job shop, energy-saving scheduling, two-mode parallel search, multi-objective discrete grey wolf optimization algorithm

摘要： 针对作业车间节能调度问题,建立了一种以优化总能耗和工件最大完工时间为目标的节能调度模型,并提出一种多目标离散灰狼优化算法进行求解。根据问题的特点,首先采用离散整数编码方式,利用调度规则生成初始种群;其次引入一种基于跟踪模式和搜寻模式的双模式并行搜索方法,并在搜索过程中动态调整两种模式下个体的数目,以协调算法全局和局部搜索能力;为了使算法适用于多目标离散调度问题,在跟踪模式下提出一种基于交叉操作的离散个体更新方法,在搜寻模式下提出一种基于记忆池机制和邻域结构的离散个体更新方法。对40个作业车间调度问题基准算例进行改造,并验证了所提算法的有效性。

关键词: 作业车间, 节能调度, 双模式并行搜索, 多目标离散灰狼优化算法