Time- and resource-based robust scheduling algorithms for multi-skilled projects

Abstract

The presence of uncertainties often disrupts the implementation of a project plan. This paper investigates robust scheduling methods for multi-skilled projects with stochastic duration. A new indicator for evaluating robustness is designed, which is used to propose a reinforcement learning-based time buffer insertion algorithm and a resource flow optimization process. An integrated time- and resource-based robust scheduling algorithm is also developed, which considers the interaction between time buffers and resource flow. Computational experiments show that the new indicator accurately evaluates the robustness of a plan, and the proposed integrated algorithm significantly outperforms traditional methods. The results demonstrate that systematically optimizing time and resources provides greater assurance that the project will proceed as planned compared to optimizing them separately.