Multi-factor based Dynamic Offloading with Coalitional Game in Mobile Cloud Computing

Distributed mobile cloud computing is a new paradigm to enhance the edge cloud computing capacity and reduce the energy of application processing. In this paper, we establish a distributed cloudlet system which consists of several cloudlets and a computation-intensive application. The application can be divided into multiple related tasks. Each task selects mobile device or the appropriate offloading cloudlet to optimize utility. Besides energy consumption, the utility includes user mobility predictive probability, channel availability and cloudlet availability. Considering these factors can overcome the problem resulting from the user mobility and reduce the probability of connection failure while ensuring processing efficiency. We formulated this dynamic offloading problem as a transferable coalition game because each task is a sensible player to maximize overall utility. Simulation results demonstrate that the dynamic offloading decision scheme we proposed can get a better offloading decision with low complexity.