An improved adaptive large neighborhood search for the home health care routing and scheduling problem with multiple mixed time windows

Abstract

The growing challenges posed by population aging and urbanization have intensified the need for efficient home health care (HHC) services to alleviate the great pressure on healthcare resources. This study addresses the Home Health Care Routing and Scheduling Problem (HHCRSP), which involves optimizing caregivers’ daily schedules while considering complex real-world constraints, including skill matching, mixed hard and soft time windows, synchronized services, and workload balancing. To address these challenges, a novel mixed-integer linear programming (MILP) model and an improved adaptive large neighbourhood search (IALNS) algorithm are proposed. The algorithm integrates an elite archive mechanism and introduces new removal and insertion operators, thus maintains archive diversity, and effectively explores the solution space through reconstruction, crossover, and mutation. Furthermore, it adopts a two-stage approach to ensure solution feasibility. Extensive computational experiments show the effectiveness of the proposed method and the competitiveness of the IALNS. Also, we examine the impact of the proportion of clients purchasing on-time services, time window penalty coefficients, and the number of available time windows on scheduling solutions. These findings underscore the proposed algorithm’s potential to reduce the cost of HHC services.