- Journal
- APPLIED SOFT COMPUTING
- Année
- 2026
- Volume
- 192
- Article
- 114779
- Mois
- APR
- DOI
- 10.1016/j.asoc.2026.114779
Abstract
Home health care routing and scheduling problem (HHCRSP) constitutes a classic combinatorial optimization problem (COP) in home health care (HHC) delivery systems. Considering the uncertain health conditions of patients and time-dependent traffic congestion in the urban area, this study investigates a real-life HHCRSP involving fuzzy service time and time-dependent travel time features. Notably, we introduce the novel considera tion of caregiver physical load, which occurs when the caregivers have to carry the care equipment by taking the stairs. The studied HHCRSP particularly emphasizes load computation for synchronized visits, where cooperative caregiving enables physical load alleviation. The problem is formulated as a fuzzy mixed-integer programming model and is solved by a customized Q-learning-based algorithm. The algorithm employs large neighborhood search (LNS) to guide the exploration and variable neighborhood descent (VND) to intensify the search ability. The customized heuristics which take advantage of the problem's features to enhance the algorithm's search abil ities are designed. Meanwhile, the removal, insertion and local search heuristics involved in the LNS and VND are selected by a Q-learning agent during the iterative process. The extensive experiments highlight the superior performance of the proposed algorithm. Sensitivity analyses highlight that the fuzzy model outperforms its deter ministic counterpart in reliability and flexibility, with its key characteristics directly governing solution quality and computational cost. Finally, empirical results confirm the practical applicability of the proposed model and algorithm. These findings make significant theoretical contributions to the HHCRSP literature by introducing the critical dimension of caregiver's physical load, while providing actionable implications for optimizing decision support systems in HHC operations. Furthermore, this study establishes a methodological framework for solving analogous COPs in urban logistics systems, with specific emphasis on last-mile delivery challenges associated with heavy-item (e.g., large furniture and appliances) transportation that entails significant physical load.