同济大学 机械与能源工程学院， 上海 201804
出版日期:2018-07-28发布日期:2018-07-28通讯作者:陆志强，男，教授，博士生导师，电话(Tel.): 021- 69589598; E-mail: zhiqianglu@tongji.edu.cn.基金资助:国家自然科学基金资助项目（61473211，71171130）Integrated Optimization of Project Splitting and Resource Investment Project Scheduling
ZONG Baoshi,LU ZhiqiangSchool of Mechanical Engineering, Tongji University, Shanghai 201804, China
Online:2018-07-28Published:2018-07-28摘要/Abstract
摘要： 基于一类实际生产决策需求，提出了依赖于项目拆分的资源投入调度问题.在分析项目拆分对资源投入影响的基础上，以资源投入最小化为目标，建立了项目拆分与资源投入调度问题的集成优化模型.结合项目拆分和资源投入调度的特点，提出了包含项目拆分优化和资源投入调度优化的两阶段集成优化算法.项目拆分阶段分析可行的拆分条件，采用项目初始拆分及局部调整的方法，可以快速获得较优的项目拆分方案.资源投入调度阶段以采用作业优先级和资源能力双列表编码的遗传算法为搜索框架，充分利用迭代过程中的信息，设计一种基于概率分布的资源能力选择方法来改进资源列表，使资源能力列表加速向最优解收敛.应用PSPLIB标准算例进行数据实验，结果证明了该算法的有效性和可靠性.
关键词: 项目拆分, 资源投入, 集成优化, 遗传算法
Abstract: Based on a class of actual production decision requirements, a resource investment scheduling problem that depends on project splitting is proposed. Based on the analysis of project splitting's impact on resource investment, the integrated optimization model of project splitting and resource investment scheduling problem is established with the aim of minimizing resource investment. A two-stage integrated optimization algorithm is proposed including project splitting optimization and resource investment scheduling optimization. The feasibility of split conditions is studied in the project splitting stage. Using an initial split and a local adjustment method, a project split program can be quickly obtained. Genetic algorithm with job priority and resource capacity dual list are used for the search frame in the resource investment scheduling stage. With full use of the information in the iterative process, a resource allocation method based on probability distribution is designed to accelerate the algorithm convergence. Results from numerical experiments using PSPLIB standard library prove the effectiveness and reliability of the algorithm.
Key words: project splitting, resource investment, integrated optimization, genetic algorithm
PDF全文下载地址:
点我下载PDF
