Research Article
Dynamic Resource Allocation and Streaming in Mobile Edges: A Deep Reinforcement Learning Approach
@INPROCEEDINGS{10.1007/978-3-030-66785-6_19, author={Daud Khan and Zeeshan Pervaiz}, title={Dynamic Resource Allocation and Streaming in Mobile Edges: A Deep Reinforcement Learning Approach}, proceedings={Machine Learning and Intelligent Communications. 5th International Conference, MLICOM 2020, Shenzhen, China, September 26-27, 2020, Proceedings}, proceedings_a={MLICOM}, year={2021}, month={1}, keywords={Mobile Edge Computing Resource allocation Deep Reinforcement Learning Real-time Double Deep Q-network}, doi={10.1007/978-3-030-66785-6_19} }- Daud Khan
Zeeshan Pervaiz
Year: 2021
Dynamic Resource Allocation and Streaming in Mobile Edges: A Deep Reinforcement Learning Approach
MLICOM
Springer
DOI: 10.1007/978-3-030-66785-6_19
Abstract
Real-Time wireless communication devices with restricted assets face more extreme limit limitations than at any other time expansion of various sophisticated and calculation severe, versatile applications. In this paper, we explore the issue of resource allocation and also use real-time devices in mobile edge networks, efficient streaming with Double Deep Q Reinforcement Learning. The ideal arrangement considering the elements the system is strict about accomplishing. We aim to develop a smart agent to improve the allocation of resources in the decision-making process. We present a new combination of double Q-learning and DQN dueling algorithms and design suggested a solution to this issue based on the Double Deep Reinforcement Learning. We implement Double Deep Q Reinforcement Learning that can take into consideration a long-term task and learn from experience. The current proposal also measures the time-varying tasks of MEC servers and discusses the strategy of transferring tasks from one to another MEC server, better optimizing the value of the task by reducing unnecessary waiting times for queue. Results from the simulation show that our proposed approach significantly decreases system costs relative to the other parameters.
