Heuristic Methods for Dynamic Multiproject Scheduling Problems with Stochastic Transfer Times

Abstract

In practice, activities of the construction multiproject use limited resources in different locations, which requires resources to be transferred frequently between projects or activities. In this study, we simultaneously consider the stochastic resource transfer times and the dynamic arrival of new projects during the multiproject execution and propose the resource-constrained dynamic multiproject scheduling problem with stochastic transfer times (RCDMPSP-STT). Since priority rule (PR)-based heuristic methods can efficiently solve project scheduling problems through real-time decision making in uncertain environments, we evaluate the performance of 72 PR-based heuristic methods based on six objectives from the project, multiproject, and resource perspectives. Each PR-based heuristic method consists of an activity priority rule (APR) and a transfer priority rule (TPR) to determine the sequence of activities and resource transfer. Moreover, based on the resource transfer information, we design eight new APRs. In addition, a hybrid PR-based heuristic method is proposed to select different PR-based heuristic methods for multiproject scheduling based on the arrival time of the new projects. Extensive numerical experiments show that the newly designed resource transfer information-based APRs and composite APRs perform better on the robustness and makespan objectives, respectively, with an improvement of 30.63% and 22.24% compared with the classical APRs. Furthermore, under different objectives and perspectives, better-performing PR-based heuristic methods are recommended for the project and multiproject managers.