Effective Strategies for the Teamwork Scheduling Problem with Job-Person Constraint

The Teamwork Scheduling Problem with Job-Person constraint (TWSPwJP) is a challenging optimization problem that involves assigning jobs to teams of workers while considering the availability of each worker and the constraints of job assignments. Due to its NP-hard complexity, solving the TWSPwJP optimally for large instances is computationally impractical. In this paper, we propose and investigate the performance of four diverse algorithmic approaches: a heuristic First-Come-First-Served (FCFS) method, two metaheuristics - Simulated Annealing (SA) and Genetic Algorithm (GA), and a learning heuristic leveraging Deep Q-Networks (DQN). This study demonstrates the advantages of the metaheuristic and learning-based approaches over the basic heuristic, providing valuable insights into the strengths and limitations of the proposed methods. The results also underscore the potential of integrating reinforcement learning with optimization for complex scheduling problems.