DQN-enabled content caching and quantum ant colony-based computation offloading in MEC

Abstract

Computation-intensive and time-delay-sensitive tasks show explosive growth, which makes mobile user devices put forward higher requirements for such tasks. Mobile user devices are difficult to meet the task requirements due to limited computing capacity. To handle this situation, a mobile edge computing system requires to be constructed. Nevertheless, in the mobile edge computing system, if no popular content is cached in the edge server, users will repeatedly send requests to a remote cloud data center when requesting content, which will cause delay and backhaul link load. To handle this situation, this article plans to put forward a content caching policy. In this policy, the caching benefit in cache replacement, delay in transmission and backhaul link load is taken into account in the content cache. Then DQN is used to handle the problem to obtain the optimal content caching scheme. In addition, in the mobile edge computing environment, the computing capacity of mobile user devices is limited, so task data needs to be unloaded to the edge server for local or remote execution to improve the computing capacity. To handle the above problems, this article plans to put forward a quantum ant colony based computing offloading strategy. This strategy takes into account three aspects: delay, energy consumption and server cost. Quantum ant colony algorithm is used to handle the situation. Experimental results indicate that the content caching strategy based on DQN can effectively improve cache hit percentage, and cut down transmission delay and return link load. Quantum ant colony based computational offloading strategies can increase task completion rates, and decrease task completion delays, equipment energy consumption and total system cost.