A multi-objective comprehensive learning particle swarm optimization with a binary search-based representation scheme for bed allocation problem in general hospital

Bed allocation is a crucial issue in hospital management. This paper proposes a multi-objective comprehensive learning particle swarm optimization with a representation scheme based on binary search (BS-MOCLPSO) to deal with this problem in general hospital. The bed allocation problem (BAP) is first modeled as an M/PH/c queue. Based on the queuing theory, the mathematical forms of admission rates and bed occupancy rate is deduced for each department of the hospital. Taking the maximization of both rates as objectives, the BS-MOCLPSO generates a set of non-dominated optimal allocation decisions for the hospital manager to select. The proposed algorithm introduces a novel binary search-based representation scheme, which transforms a particle´s position into a feasible allocation scheme through binary search. Simulation results on real hospital data show that the proposed algorithm can offer allocation decisions that lead to higher service level and better resource utilization.