工业工程

工业工程

• 专题论述 • 上一篇    下一篇

基于资源限制的多目标网络计划优化模型

  

  1. 北京科技大学 1.机械工程学院； 2. 东凌经济管理学院，北京 100083
  • 出版日期:2015-12-31 发布日期:2016-03-24
  • 作者简介:钮建伟（1977-），男，河北省人，副教授，博士，主要研究方向为项目管理、网络计划技术。
  • 基金资助:

    国家自然科学基金（面上）资助项目（71172169）

Multi-objective Optimization Modeling for Resource Constrained Project Scheduling Problem

  1. 1. School of Mechanical Engineering, University of Science and Technology Beijing, Beijing 100083, China;2. Dongling School of Economics and Management, University of Science and Technology Beijing, Beijing 100083, China
  • Online:2015-12-31 Published:2016-03-24

摘要： 目前网络计划优化研究要么没有考虑资源限定的柔性，要么只是集中于单纯的工期优化或资源优化等单目标优化。本文针对传统网络计划建模资源限制缺少柔性、优化目标单一等问题进行了深入的研究。在柔性资源的限制下，为使得工程网络计划达到总体最优，考虑工程项目的工期、成本、项目净现值、资源的均衡等多个目标，建立其网络计划优化模型，并采用粒子群算法予以求解。根据拓扑排序算法生成满足时序约束的活动序列并计算活动的时间参数。对于产生资源冲突的活动，依照执行优先权解决冲突资源的执行顺序，更新时间参数。采用随机权重的方法，让粒子群算法种群的多个个体进行随机转化，从而保持解的多样性。采用国际上通用的Patterson问题库中benchmark算例对本文提出的方法进行验证。结果表明，与初始方案相比，优化后的方案分别在工期上缩减了20%，成本上缩减了11.17%，净现值增加了11.82%，资源均衡度减少了18.29%。由此可见，提出的基于粒子群算法的优化模型对资源限制下的网络计划中的工期、成本、净现值、资源均衡度等多个目标均实现了不同程度的优化。

关键词: 网络计划, 多目标优化, 柔性资源, 建模, 粒子群算法

Abstract:

Recent studies on project scheduling problem are usually considered as lacking resource constraint issue or just concentrat on makespan or resource utilization optimization. Considering the disadvantages of traditional modeling methods, such as lacking flexibility or owning solely a single optimization objective, a novel approach is presented to optimize the problem of resource-constrained multi-objective optimization in project scheduling. The objective is to minimize the makespan and cost, maximize the net present value and level the resource fluctuation as well. An initial project schedule is originally constructed based on topology sorting algorithm, which satisfies the time sequence constraint. As for those activities in conflict with each other in resource utilization, priority rule is adopted to solve the conflicts. A case study withdrawn from Patterson benchmark database is conducted to illustrate the effectiveness of this method. Particle swarm optimization algorithm is adopted to realize the optimization. During the calculation, random weight is used to make sure the randomization of particle swarm evolution, which prevents the algorithm from trapping into local optimization. Results show that, compared with the original solution, 20% makespan reduction, 11.17% cost reduction, 11.82% increase of the net present value, and 18.29% reduction of resource fluctuation are obtained based on the proposed approach. This research shows the ability of the proposed method in optimization of project scheduling considering makespan, cost, net present value and resource fluctuation as well.

Key words: network planning, multi-objective optimization, flexible resource, modeling, particle swarm optimization algorithm