Home Care Scheduling Considering Preference Matching

This paper addresses a home care scheduling problem faced by home care service providers.Home care scheduling is an extension of vehicle routing problem with specific constraints and characteristics,like preference matching,continuity of care and temporal dependency.As the demand for home care services increases,the competition in the market becomes increasingly fierce.Currently,most service providers do the scheduling work manually,which may result in high organizational efforts and high costs.Therefore,to do scheduling work efficiently while decreasing cost and increasing service quality becomes quite important for providers.This research mainly focuses on three parts:Firstly,based on the general characteristics and constraints of home care scheduling,a general home care scheduling problem considering preference matching is studied.A mathematical model is built and a variable neighborhood search algorithm is proposed,with neighborhood operators and shaking procedure designed.The algorithm is tested on several datasets and numerical results show that the proposed algorithm is effective.Secondly,considering the detailed characteristics of the real scheduling problem faced by a home care service provider,like continuity of care,night shifts and clients’ personal preference for caregivers,a real case home care scheduling problem is studied.A mathematical model is built and the variable neighborhood search algorithm is applied,with process to generate initial solutions,neighborhood operators and shaking procedure designed for this problem.The algorithm is tested on datasets from real operations of the provider and numerical results show that the proposed algorithm can efficiently provide high quality solutions.Finally,considering the multiple objectives of the home care scheduling problem,including minimizing travel time,overtime cost,preference mismatches and unsatisfied continuity of care constraint,a home care scheduling problem considering multiple objectives is studied.Multiobjective formulation is presented and the epsilon-constraint method is applied to solve the problem.Numerical experiments are conducted with several datasets and the results are quite helpful for the real operations of the provider.