Computation Offloading in Heterogeneous Mobile Edge Computing With Energy Harvesting

Energy harvesting aided mobile edge computing (MEC) has gained much attention for its widespread application in the computation-intensive, latency-sensitive and energy-hungry scenario. Computation offloading, which leverages powerful MEC servers (MEC-ss) to augment capability of less devices (MDs), is intrinsically a distributed over heterogeneous networks. In this article, computation offloading from multi-MD multi-MEC-s systems with energy investigated game theoretic perspective. The objective minimize average response time an MD that consists communication time, waiting processing time. M/G/1 queueing models are established MDs' generation MEC-ss' task receiving. interference among MDs, randomness generation, harvested arrival, wireless channel state, at MEC-s, power budget constraint taken into consideration. A noncooperative formulated. action vector denotes amount tasks offloaded all MEC-ss (the element value can be zero) local process. We give definition existence analysis Nash equilibrium (NE). Furthermore, we reconstruct optimization problem MD. 2-step decomposition presented performed. Thereby, arrive one-dimensional search greatly shrunken sub-problem. sub-problem nonconvex, but Karush-Kuhn-Tucker (KKT) conditions have finite solutions. obtain optimal solution by seeking Thereafter, distributive NE-orienting iterated best-response algorithm designed. Simulations carried out illustrate convergence performance effectiveness proposed algorithm.