A Decision Support System for daily scheduling and routing of home healthcare workers with a lunch break consideration

This study examines a home healthcare scheduling and routing problem (HHSRP) with a lunch break requirement. This problem especially consists of lunch break constraints for caregivers in addition to other typical features of the HHSRP in literature such as hard time window constraints for both patients and caregivers and patient preferences. The objective is to minimize both travel distance in a route and unvisited patient (penalty) cost. For this NP-Hard problem, we developed an effective Adaptive Large Neighborhood Search algorithm to provide high-quality solutions in a short amount of time. We tested the proposed four variants of the algorithm with the selected problem instances from the literature. The algorithms provided nearly all optimal solutions for 30-patient problem instances in 12 seconds on average. Additionally, they provided better solutions to 36 problem instances up to 36% improvement in some instance classes. Moreover, the improved solutions achieved to visit up to 10 more patients. The algorithms are also shown to be very robust due to their low coefficient variance of 0.3 on average. The algorithm also requires a very reasonable amount of time to generate solutions up to 54 seconds for solving 100-patient instances. A decision support system, namely Home Healthcare Decision Support System (HHCSS) was also designed to play a positive role in preventing the COVID-19 global pandemic. The system employs the proposed ALNS algorithm to solve various instances of approximately generated COVID-19 patient data from Turkey. The main aim of developing HHCSS is to support the administrative staff of home healthcare from the tedious task of scheduling and routing of caregivers and to increase service responsiveness.