Interference Mitigation in B5G Networks: A Critical Review of Conventional, Optimization-Based, and Ai-Driven Techniques

Abstract

Beyond fifth generation (B5G) networks require effective interference mitigation to maintain high throughput, low latency, spectral efficiency, and reliable communication in dense and dynamic environments. This paper presented a structured critical review of B5G interference-mitigation techniques. Twenty-five studies were selected from major scientific databases based on clearly defined interference problems, mitigation methods, implementation approaches, and measurable performance outcomes. The techniques were grouped into conventional and coordination-based, optimization-based, and artificial-intelligence-driven approaches. Conventional methods such as power control, fractional frequency reuse, coordinated scheduling, coordinated multipoint transmission, and beamforming remain effective but may lack adaptability under rapidly changing conditions. Optimization-based methods provide systematic resource control but are limited by computational complexity and dependence on network-state information. AI-driven approaches, particularly reinforcement and deep reinforcement learning, offer greater adaptability but face challenges involving training, convergence, computational demand, and signaling overhead. Overall, no single technique is universally suitable; effective B5G interference management requires balancing performance gains with complexity, scalability, overhead, and practical implementation.