Multi-Leader–Follower Game for MEC-Assisted Fusion-Based Vehicle On-Road Analysis

Abstract

Multi-source data fusion to support vehicle on-road analysis is a promising service offered by mobile edge computing (MEC) for vehicles. With the fusion results delivered in near real-time, vehicle users (VUs) can peek around the corner, extend sensing range, reinforce and validate local observations. Consequently, there has emerged a new market between smart mobility service providers (SMSPs) and VUs in offering and purchasing multi-source data fusion results to support vehicle on-road analysis. Each SMSP and each VU compete with their peers to maximize their own profits. Also, VUs can receive multiple fusion results from several SMSPs and combine them to achieve a better inference. In this paper, we develop a multi-leader-follower game to model this complicated coupled problem. For the single-SMSP scenario, we analyze the properties of the leader-follower (L/F) Nash equilibrium and then reformulate the game as a mathematical program with equilibrium constraints (MPEC) to obtain the equilibrium. For the multi-SMSP scenario, the game is reinterpreted as an equilibrium problem with equilibrium constraints (EPEC), for which we analyze the local Nash equilibrium (LNE) with the assistance of variational inequality (VI) theory. Then, the block coordinate descent (BCD) method, which is low-complexity, is applied to solve the EPEC. Finally, numerical results are provided to validate the theoretical analysis and show that our proposed strategies maximize the utilities for both SMSPs and VUs.