带有活动重叠的资源受限项目调度问题建模与求解

在新产品研发项目中,常常通过活动重叠缩短工期,但是重叠会带来返工风险,进而造成额外的时间和资源的消耗.将活动重叠分为自然重叠和强制重叠,采用设计结构矩阵、紧前重叠活动时间因子矩阵、紧后重叠活动时间因子矩阵和重叠返工影响矩阵量化重叠返工对时间和资源的影响.以最小化研发项目工期为目标,建立了带有活动重叠的资源受限项目调度问题优化模型,并设计了改进的遗传算法进行求解.该算法采用基于优先规则的编码机制,并将考虑活动重叠的进度生成机制作为解码方法.实例研究结果表明,提出的考虑活动重叠的模型与算法,不仅可以缩短项目工期,而且同时使项目资源的使用更加均衡.最后在随机生成的算例集合上测试了算法的有效性.

Modeling and solving the resource-constrained project scheduling problem with activities overlapping

The activity overlapping in new product R&D project is an important approach to shorten project duration, but overlapping might lead to rework risk and take additional time and resources consumption. Therefore, the dependency structure matrix (DSM), preceding overlapping activity time factor DSM, succeeding overlapping activity time factor DSM and overlap impact matrix are proposed respectively to describe the natural overlapping, forced overlapping and the influence of duration and resource due to overlapping rework. A resource-constrained project scheduling optimization model with overlapping activities is built, and an improved genetic algorithm is presented to solve the model based on priority rules and schedule generation mechanism with overlapping. An instance is studied and the results indicate that the model not only can shorten project duration, but also realize the balance of resources. Finally, application of the algorithm to examples of different sizes and nature demonstrates the validity of the algorithm.