中文摘要：

在多平行工作站环境下，为使限定资源分配下的车间调度问题（JobShopproblem，JSP）具有最小总延迟时间；同时又可设定各订单具有不同的开工日（releasedate）及到期日，提出以可开工时间与结束时间为基础的分解解法，并在遗传算法的基础上构造混合遗传算法（hybridgeneticalgorithm，HGA）来实现目标设定。实验结果表明，HGA在问题求解质量与Lingo解的最佳解差异在15％以内，并具备较基本型遗传算法更佳的稳定性。结果显示该算法可帮助管理人员实现智能资源配置与订单调度。

英文摘要：

This study addressed a job scheduling and resource allocation problem with distinct release dates and due dates to minimize total tardiness in parallel work centers with a multi-processor environment. To solve the problem, this study also pro- posed a hybrid genetic algorithm （HGA） with release and due dates based decomposition heuristic. Experimental results show that the percentage deviations between the HGA and Lingo are smaller than 15%, and the HGA has smaller variance than the GA. This study proposed a decision-supporting model, which integrated simulation, genetic algorithms and decision support tools, for solving the JSRA problem by practical perspective.