Multi-Criteria Handover in Optical Wireless and Radio Frequency Heterogeneous Networks

This study tackles the handover challenge in optical wireless communication (OWC)/radio frequency (RF) heterogeneous networks (HetNets) by introducing two complementary and novel approaches. First, we present OPHO, an optimization framework formulated as a multi-objective mixed-integer linear programming (MILP) model. OPHO simultaneously minimizes the number of active access points (APs), node-to-AP distances, and handover costs, while accommodating various handover types. To balance conflicting objectives, goal programming is employed. Although OPHO provides a theoretically optimal benchmark, its centralized architecture and computational complexity hinder scalability and adaptability in dynamic environments. To address these limitations, we propose MAHO, a decentralized multi-attribute decision-making (MADM) scheme that combines a modified Kalman filter with fuzzy TOPSIS. Unlike conventional MADM-based methods, MAHO incorporates multiple predicted states—generated via the modified Kalman filter—as distinct decision attributes. This predictive multi-horizon modeling embeds forward-looking network conditions directly into the decision-making process. Furthermore, fuzzy TOPSIS serves as the decision-making engine, effectively handling uncertainty in both current and predicted metrics. Simulation results show that while OPHO achieves optimal handover performance, MAHO delivers near-optimal results with significantly lower computational overhead, making it highly suitable for real-time deployment in dynamic and resource-constrained OWC/RF HetNets.