Abstract: Resource breakdown occurs frequently during the implementation of projects. It may lead to the changes of project schedule and generate additional expenses. This paper involves the project scheduling problem under resource breakdown, where the objective is to adjust the baseline schedule reasonably so as to minimize the incurred additional expenses. The problem is identified at first and the optimization model is constructed accordingly. For the NP-hardness of the problem, a tabu search heuristic algorithm is developed. Finally, given the baseline list algorithm and the random generation algorithm as comparison, we test the tabu search algorithm on a set of standard instances generated randomly. The conclusions are drawn as follows. First, within the acceptable computation time, the quality of the desirable solutions obtained by the tabu search heuristic algorithm is significantly better than those obtained by other two heuristic algorithms. Second, the average computation time increases with the activity number, but decreases with the network complexity, the renewable resource strength, and the number of resource breakdown, respectively. Third, the mean of objective function value also climbs with the activity number and drops with the network complexity and the number of resource breakdown, but it seems that there is no significance influence on the renewable resource strength.

Key words: reactive project scheduling, optimization model, tabu search, stochastic resource breakdown

摘要： 资源中断是项目实施过程中一种常见现象,它会导致项目进度计划的变更并引起额外的成本。本文研究资源随机中断下的项目调度问题,目标是对基准进度计划进行合理的调整,以最小化由此所造成的额外成本。作者首先对研究问题进行界定,随后构建问题的优化模型。针对模型的NP-hard属性,设计禁忌搜索启发式算法。最后以基准列表算法和随机生成算法为参照,在随机生成的标准算例集合上对算法进行测试,得到如下结论:在可接受的计算时间范围内,禁忌搜索获得的满意解质量明显高于其他两种启发式算法;算法的平均计算时间随着项目活动数的增加而增加,随着网络复杂度、资源强度或资源中断次数的增加而减小;满意解的平均目标函数值,随着项目活动数或网络复杂度的增加而增加,随着资源中断次数的增加而减小,与资源强度无明显关系。

关键词: 反应性项目调度, 优化模型, 禁忌搜索, 资源随机中断