KBGWO-RNP: Knowledge-Based Grey Wolf Optimizer for Multi-Criteria RFID Network Planning in Medical Asset Monitoring

Radio Frequency Identification (RFID) has emerged as an effective remote technology for real-time monitoring and management of medical assets in hospitals. Most existing RFID Network Planning (RNP) methods are primarily based on either heuristic or metaheuristic approaches. While heuristic approaches are computationally efficient and converge rapidly, they often suffer from premature convergence and suboptimal network configurations. Conversely, metaheuristic algorithms provide stronger global search capabilities and improved solution quality, but they typically require higher computational effort and may still exhibit stagnation in local optima when applied to complex hospital layouts. To overcome these limitations while utilizing the strengths of both paradigms, this paper proposes a Knowledge-Based Grey Wolf Optimizer for RNP, referred to as KBGWO-RNP. The proposed method integrates the global exploration capability of the metaheuristic-driven search with knowledge-based heuristic operators that guide local search and refinement. In particular, the framework incorporates domain-specific knowledge to enhance antenna placement decisions and improve convergence behavior. The KBGWO-RNP framework supports directional antennas with varying coverage profiles. A multi-criteria objective function is formulated to increase the network coverage while simultaneously reducing inter-antenna interference and deployment cost. Extensive simulation experiments conducted on a hospital layout demonstrate that the proposed KBGWO-RNP framework consistently outperforms conventional heuristic and metaheuristic baselines. The results show that the proposed method achieves a coverage rate of 90.4% while maintaining the interference level at 19.9%, indicating a strong balance between performance different objectives. Furthermore, ablation analysis confirms that the integration of knowledge-based guidance with metaheuristic search significantly improves both solution quality and stability. The proposed framework offers a balanced trade-off between computational efficiency and optimization performance, and demonstrating clear advantages over existing approaches.