Game Theoretic Resource Planning and Request Scheduling in Mobile Edge Computing Networks

Mobile Edge Computing (MEC) networks offer an increasing computing power through the collaboration among MEC nodes. This opens a large computing market and brings challenges for efficient resource management. In this paper, we study the joint optimization problem of planning cost-efficient edge networks, allocating link and computation resources, as well as scheduling and routing user requests in edge computing networks with arbitrary topologies and multiple ingress nodes. We formulate this problem as a Stackelberg game where the network operator, as the leader, aims at maximizing its profit, and the edge nodes, as the followers, minimize their users' costs and latency. Then, we prove the existence of the generalized Nash equilibrium for the follower subgame, and the Stackelberg equilibrium for the leader-follower game. We further propose a distributed best-response algorithm for the follower game and an alternating leader-follower optimization algorithm for the full game to compute the equilibrium and prove its convergence. A centralized optimization incorporating both profit and network latency targets is formulated and solved, which serves as benchmark for the game solution. Extensive numerical results demonstrate the effectiveness of the proposed game, achieving near-optimal planning and scheduling solutions in a very short time even for large-scale edge networks.