A Multi-objective Reinforcement Learning Solution for Handover Optimization in URLLC

The growth of wireless communications facilitates advanced technologies like Tactile Internet and robotics, requiring ultra-reliable low-latency communications (URLLC). Effective user equipment (UE) handover between access points (AP) is crucial for URRLC to enhance Quality of Experience (QoE). However, HO failures can occur due to various reasons, impacting URRLC use cases reliant on service reliability. This paper introduces HORLA, a multi-objective reinforcement learning model that enhances received signal strength and reduces outage probability during AP selection for UE. HORLA surpasses the conventional HO algorithm, reducing failure attempts by over 40% and cutting energy consumption for reattempted HO requests by nearly 57%.