Design And Implementation Of Machine Learning Based Aadptive Resource Scheduling Algorithm For Future Communication Networks

In order to meet the requirement of ultra high speed,throughput,reliability,low latency metrics in the future communication network,the emerging network architecture will be more and more heterogeneous,hyper-intensive and diversified,which will cause the Inter-Carrier Interference(ICI)more and more multi-level and complex.The issues of intense ICI and resource collision will greatly limit the capacity of the network.Therefore,these emerging networks require a reasonable and effective resource allocation strategy,which can reduce the execution time of the task,improve service quality and increase the whole network throughput and overall resources utilization rate of the system.However,the conventional resource scheduling strategy cannot cope with this super dense network structure.What's more,adapting to changes of the environment is another big challenge.In order to make full use of the network's own conditions and system environment changes,it is necessary to design a self-organizing and self-configuring resource scheduling technology with some intelligent algorithm.As a typical algorithm of machine learning,Q-learning algorithm performs well in helping the agents to get the best experience by self-learning.The main purpose of this paper is to propose an adaptive resource allocation strategy for some major future communication networks with machine learning algorithm.In this paper,we have studied the application of machine learning algorithms from two directions,the first one is direct resource scheduling decisionsn and the other is indirect system parameter selection.Besides,we make some comparison on different learning subjects and different learning methods.Finally,we show the simulation results and make the conclusion analysis,which proves the effectiveness of the proposed resource-scheduling algorithms.