An adaptive-learning-based genetic algorithm for collaborative scheduling of distributed operating rooms

Abstract

To improve surgical services and hospital performance, collaborative operating room planning and scheduling across a network of hospitals has recently emerged as a new challenge in both healthcare industry and academic community. The paper considers a novel distributed operating room scheduling problem (DORSP), in which elective patients planned on a given day are scheduled for surgeries in the distributed operating rooms of collaborative hospitals. Since hospitals are different in terms of specialization and medical expertise, some complicated surgeries in this problem can only be performed in a subset of collaborative hospitals. Based on the similarities between healthcare delivery systems and production systems, DORSP is modelled as a distributed two-stage no-wait hybrid flow shop scheduling problem with factory eligibility. To deal with the NP-hardness of DORSP, an adaptive-learning-based genetic algorithm (ALBGA) is proposed to generate collaborative surgery schedules. In addition to traditional genetic operators, ALBGA also applies an adaptive learning operator to enhance the search ability by mimicking human learning behaviours. Computational results on both small-sized and large-sized test problems show that ALBGA is competitive among the compared algorithms.