ECS-NBS: Exact Computation of Sequential Nash Bargaining Solutions

Abstract

In order to support reliability and safety in autonomous driving services, vehicular networks should be able to efficiently and fairly allocate the time-varying and limited resources to multiple vehicles as quickly as possible. In this paper, we adopt the Nash bargaining solution (NBS) as the resource management strategy. However, the resource allocation based on the NBS requires exponentially increasing computational complexity for dynamically changing resources over time. We propose an analytic solution, referred to as the exact computation of the sequential NBS (ECS-NBS), to perfectly compute the NBS sequentially without any iterations. The key idea is to use the axiom of independence of linear transformations in NBS for the design of a transformation matrix that captures the changes in the adjacent feasible utility sets. This enables the NBS to be invariant to the change of utility sets over time, yielding the perfectly accurate NBS with the lowest complexity, as confirmed through simulations.