Optimizing Referee Assignments with Fuzzy Evaluation and Particle Swarm Intelligence

Referee assignment is a recurring challenge in a wide range of settings—from assigning officials to sports matches to allocating reviewers for academic conferences. These scenarios often involve complex constraints, such as assigning multiple referees to a single activity or allowing referees to evaluate multiple tasks. In this study, we address a generalized version of the referee assignment problem, focusing on the case where each activity is matched with exactly one referee. To evaluate the suitability of assignments, we leverage rating scores across multiple dimensions for both referees and activities. These scores are incorporated into fuzzy membership functions that quantify assignment quality. We propose a Particle Swarm Optimization (PSO) algorithm that uses these fuzzy-based evaluations to guide the search for optimal assignments. Synthetic datasets with varying levels of data dispersion—generated from uniform distributions with equal means but different variances—are used to test the algorithm. Our experimental results demonstrate that the proposed PSO approach is not only computationally efficient but also robust across different variance scenarios. This method offers a flexible and adaptable solution for a broad class of referee-assignment problems.