关键词: 多项目调度, 净现值最大化, 优化模型, 变邻域算法, 非共享资源

Abstract: Taking the maximization of the net present value (NPV) as the objective, this paper investigates the multi-project scheduling problems with non-shared resource constraints. In the problem, the contractor has to allocate resources to individual projects and then under the constraint of the allocated resources, the schedule of each project is arranged independently to maximize its NPV. First, the practical and theoretical backgrounds of the multi-project scheduling problem with non-shared resource constraints are introduced, the research problem is defined, and its research significance is demonstrated. Then, based on the notation definition, the multi-project scheduling optimization model consisting of upper and lower sub-models is developed, and the three basic properties of the problem are proposed. Afterwards, in light of the characteristics of the studied problem, a two-module nested variable neighborhood search heuristic algorithm is designed, where the proposed properties are integrated to enhance the searching efficiency of the algorithm. Finally, in randomly generated standard instances, a large-scale computational experiment is conducted to evaluate the performance of the designed algorithm and analyze the effects of key parameters on the objective function. The findings demonstrate that among the four algorithms compared in the experiment, the variable neighborhood algorithm designed in this paper is the most promising algorithm for the studied problem. The NPV of projects ascends with the increase in the milestone activity number, advanced payment proportion, middle payment proportion, and project deadline while descends with the increase in the cash flow discount rate and resource factor.

Key words: multi-project scheduling, net present value (NPV) maximization, optimization model, variable neighborhood algorithm, non-shared resource