Incentive-Based Game-Theoretic Framework for Sustainable 5/6G Cellular Networks

A Deepened Two-Stage Stackelberg Analysis with Multi-Scenario Validation

Abstract

The explosive uptake of data-hungry services—from 8 K video streaming to XR and massive-scale IoT—pushes 5/6G cellular networks to their engineering and ecological limits. Wi-Fi–centric mobile-data offloading (MDO) is a low-capex remedy, yet existing schemes seldom balance (i) economic fairness between Mobile Network Operators (MNOs) and third-party Access Points (APs), (ii) fine-grained traffic prioritisation, and (iii) verifiable sustainability gains. We extend earlier work by tripling analytical depth and experimental breadth:
1. A comprehensive heterogeneous triple-tier model (macro–micro–Wi-Fi) that embeds energy, carbon, and monetary components.
2. A two-stage Stackelberg game in which the MNO (leader) jointly sets spectrum price, energy rebate, and traffic-type incentive, while APs (followers) optimise load, power level, and admission control.
3. Rigorous proofs of existence, uniqueness, and Pareto optimality of equilibrium, plus a repeated-game extension that guarantees long-run coalition stability.
4. Two distributed algorithms—Adaptive Best Response (ABR) and Primal-Dual Incentive Descent (PDID)—with O(n) and O(log n) message complexity, respectively.
5. A MATLAB/ns-3 co-simulation over five realistic scenarios (Dense-Urban, Suburban, Rural, Campus, Stadium) fed with 3GPP TR 38.901 channel traces, renewable-energy price curves, and Cisco VNI 2024 traffic forecasts.
Results show that, relative to four state-of-the-art baselines (IFPC, IMDO, RAIM, and LSCCOA), our frame-work
• raises the mean offload ratio by 42 %,
• boosts aggregate downlink throughput by 33 %,
• cuts average delay by 29 %,
• lowers per-GB carbon intensity by 37 %, and
• improves the Jain fairness index for profit sharing from 0.78 to 0.94.