Optimization Method Study for Scheduling of Physicians in Fever Clinics under COVID-19 Combined with Machine Learning
To address the physician scheduling problem in fever clinics under the COVID-19 epidemic,the joint scheduling problem of physicians in the common and special fever clinics was considered.A method was proposed to calculate the patient queue length of a time-varying,stochastic system based on the queueing theory and stationary fluid approximation.Consider different constraints on scheduling in the two clinics,a two-stage algorithm was designed to solve this scheduling problem,which determined each physician's working time in one week.The first stage,a mixed integer programming model was constructed to determine the number of required doctors per hour.The model was solved efficiently by the commercial solver combined with cut planes generated by machine learning.The second stage,a branch and price algorithm was designed to solve the scheduling problem exactly and obtained a physician schedule that satisfied the requirement.Numerical experiments show that machine learning can significantly reduce the solving time of the model for the first stage.The scheduling obtained by the algorithm can effectively control the number of waiting patients and reduce the total physician working time compared to realistic scheduling.
Covid-19 pandemictime-varying queueing systemphysician scheduling problemmachine learningbranch and price algorithm
万方数据
