Shortest Path Problem for Resource-Constrained Project Scheduling Based on Deep Reinforcement Learning

At present, most scheduling problems of resource-constrained projects are static scheduling, and there are few studies on dynamic scheduling of resource-constrained projects with random construction periods. Therefore, establish a dual-objective optimization model to minimize the project duration and resource constraints, design the corresponding agent interaction environment Markov decision process is used for solving the problem of too large with dispersed state and action space. The experiment is applied to small-scale project scheduling and large-scale project scheduling cases. The optimization scheme of project shortest path scheduling is finally generated.The experiment show that resource-constrained project scheduling based on deep reinforcement learning can effectively shorten the project duration.