Reinforcement Learning for Resource Constrained Project Scheduling Problem with Activity Iterations and Crashing

Abstract Resource allocation is a key decision-making process in project management that assigns resources to activities of a project and determines the timing of the allocation in a cost and time effective manner. In this research, we address the resource allocation for a project, where iterations between activities of the project exist and the crashing, a method to shorten the duration of an activity by incorporating additional resources, is available. Considering the stochastic nature of project execution, we formulate the resource allocation as a Markov decision process and seek the best resource allocation policy using a deep reinforcement learning algorithm. The feasibility and performance of applying the algorithm to the resource allocation is then investigated by comparison with heuristic rules.