An Intelligent Optimization Strategy for Medical Doctor Rostering Using Hybrid Genetic Algorithm-Particle Swarm Optimization in Malaysian Public Hospital

Comparing manual rostering to automated rostering reveals that manual rostering is typically more challenging, time-consuming, and exhausting for doctors, particularly due to shifting business regulations, a shortage of healthcare professionals, and heavy workloads. During rostering, it is essential to consider both hard and soft constraints to minimize constraint violations, maximize medical doctor satisfaction, and meet all requirements for hard constraints. To address these challenges, this paper proposes Hybrid Genetic Algorithm and Particle Swarm Optimization (Hybrid GA-PSO) to model rostering. In this approach, one set population of working days represents the rostering structure, which is determined using evolutionary-inspired operators, search, and update procedures. Additionally, the paper conducts observations and interviews with relevant personnel in a Malaysian hospital to gather insights and highlight constraints associated with medical doctors rostering. Rostering requirements determine the relative importance of the hard and soft constraints. The results of the research indicate that the Hybrid GA-PSO approach can produce workable rosters that reduce the workload of physicians and shorten the time needed to create rosters by the total violation of both soft and hard constraints and accuracy. It also ensures compliance with both hard and soft criteria and improves rostering accuracy.