Optimization of Non-Identical Queued Offloading in Mobile Edge Computing

Offloading is a promising approach in mobile edge computing (MEC) since some tasks in the end devices could be computationally demanding. In this paper, a refined optimization model is developed for the offloading paradigm. This model introduces several features which were often omitted in existing studies, including fine-grained discrete optimization, non-identical task classes in the queue of end server, non-singular CPU cycle allocation, and non-singular pathloss terms. The model is implemented in Matlab with representative parameters. Some insightful optimization results are obtained.