Quantifiable Cost-Benefit Optimization Framework for LEO Satellite-Terrestrial Cooperative IoT: Balancing Resource Consumption and User Satisfaction

Abstract

In resource-constrained low Earth orbit satellite-terrestrial cooperative Internet of Things (L-STCIoT), overloaded user demand poses severe challenges to underlying network performance optimization. To address the resource allocation problem in this scenario, a joint resource allocation framework is proposed, enabling the satellite to make coordinated decisions on user selection and transmission resource allocation based on user demands. In this framework, satellite resource cost and user-side satisfaction are defined as quantifiable metrics, respectively, and are jointly modeled to maximize the system’s weighted profit. Due to the nonconvexity of the original problem and strong variable coupling, it is decomposed into two subproblems: user selection and transmission resource allocation. Specifically, the former is addressed using a relaxation-based minorize–maximization (MM) algorithm combined with a discrete mapping (DM) mechanism, where a strongly concave surrogate function is constructed to improve decision efficiency and stability. The latter is reformulated as a linear programming (LP) problem via variable decoupling and solved using the interior-point method, thereby reducing computational complexity. In the performance analysis section, the convergence of the proposed joint optimization algorithm is proved based on the Lyapunov convergence framework, demonstrating that the performance loss introduced by the DM mechanism is bounded and that the final solution obtained by the algorithm is a globally approximate optimal solution to the original optimization problem. Simulation results demonstrate that the method proposed in this article effectively balances satellite resource cost and user-side satisfaction under different system scales and user demand distributions, while significantly improving the overall system profit, thereby exhibiting superior performance.