Deep Q-Learning based resource allocation and load balancing in a mobile edge system serving different types of user requests

Abstract With the expansion of communicative and perceptual capabilities mobile devices in recent years, number complex high computational applications has also increased rendering traditional methods traffic management resource allocation quite insufficient. Recently, edge computing (MEC) emerged as a new viable solution to these problems. It can provide additional features at network allow alleviation limit while increasing performance for critical especially terms latency. In this work, we addressed issue reducing service delay by choosing optimal path MEC network, which consists multiple servers that different capabilities, applying load balancing where requests need be handled simultaneously routing selection based on deep- Q (DQN) algorithm. A novel control method is proposed deep Q-learning (DQL) allows end-to-end cellular networks network. Real life scenarios with various types user are considered DQL scheme adaptively assigns resources proposed. The algorithm optimizes distribution between total time use available under varying environmental conditions.