Local search approaches for patient scheduling problem in parallel operating theatre

Patient scheduling in operating theatres (OT) involves high costs to hospitals and may affect treatment and satisfaction to patients. Scheduling for patients in parallel OT can help to schedule a large number of patients and emergency patients are considered in this study .Other than that, emergency patients are able to be scheduled and treated on the day as the modification of schedule is performed during the dynamic processof scheduling. The objective of the study is to propose different types of local search which are local search(LS), improved local search(ILS) and mixed local search (MLS) for solving patient scheduling problem in parallel OT. This study presents two phases of scheduling where initial schedule of regular patients is obtained in the first phase, then scheduling process continues in the second phase when operations start and emergency patients arrive. In the first phase, pre-processing stage, combination of pre-processing stage with low-level heuristic and genetic algorithm are used.The different types of LS are applied in the second phase of scheduling. Tests are conducted for the proposed methods to look for a method that has the ability to prevent delays of patients with a lower total cost. The performance of the LS is compared with zero-one programming method from literature and it is reliable to the problem. The result of ILS has reduced the average total cost to 30:4% while MLS has achieved 52:7% decrease in the average total cost. The computational results show that MLS out performs both methods of LS and ILS by minimizing the total cost of OT greatly.

Text FS 2021 16 - IR.pdf Download (1MB)

Actions (login required)

View Item