Computation offloading game for multi-task multi-helper fog networks

Fog computing has risen as an evolving architecture to support delay-sensitive applications in Internet of Things (IoT) and next generation mobile networks. For a typical heterogeneous fog network consisting of many fog nodes, some of them have different computation tasks while some have spare computation resources, which forms a multi-task multi-helper (MTMH) network. How to effectively map multiple tasks into multiple helper nodes to reduce the service delay is a key issue to be resolved. To tackle this issue, a computation offloading problem minimizing every taskâ™s delay is considered, from the perspective of individuals. This problem is further formulated into a non-cooperative game, i.e., MTMH computation offloading (MTMHCO) game, to model the competition among tasks for helpers. The existence of Nash equilibrium (NE) is guaranteed and an efficient distributed algorithm is developed to achieve an NE for the MTMHCO game. Theoretical analysis and simulation results show that the proposed algorithm can offer the nearoptimal performance in system average delay and achieve more number of beneficial task nodes, at two orders of magnitude lower complexity than a centralized optimal algorithm.