Wireless network resource allocation method based on deep reinforcement learning

Abstract

The invention provides a wireless network resource allocation method based on deep reinforcement learning. The energy efficiency in time-varying channel environment can be maximally improved with relatively low complexity. The method comprises the following steps: establishing a deep reinforcement learning model; modeling the time-varying channel environment between a base station and a user terminal as a time-varying Markov channel in a finite state, determining a normalization channel coefficient, and inputting a convolution neural network qeval, selecting an action with the maximum output return value as a decision action, and allocating the sub-carrier for the user; allocating downlink power for the user reusing on each subcarrier based on the inverse ratio of the channel coefficient according to a subcarrier allocation result, and determining a return function based on the allocated downlink power, and feeding back the return function to the deep reinforcement learning model; andtraining the convolution neural network qeval and qtarget in the deep reinforcement learning model according to the determined return function, and determining the power local optimal allocation underthe time-varying channel environment. The wireless network resource allocation method provided by the invention relates to the field of the wireless communication and artificial intelligence decision.

Description

technical field [0001] The present invention relates to the fields of wireless communication and artificial intelligence decision-making, in particular to a method for allocating wireless network resources based on deep reinforcement learning. Background technique [0002] Since the Long Term Evolution (LTE) era, the network construction architecture has changed from a macro network to a macro-micro synergy. The sustainable development of a macro cell (Macro Cell) is facing many challenges, such as unpredictable service growth requirements, ubiquitous access Incoming demand, random hotspot deployment, and the great cost pressure of the macro cell itself. Therefore, small base stations (Small Cells) such as microcells and femtocells can provide precise coverage and supplement the advantages of blind areas, and gradually become an important link in network deployment to work with macro base stations and share the service pressure of macro base stations. The fifth generation o...

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is to provide a wireless network resource allocation method based on deep enhanced learning to solve the problem in the prior art that wireless resource allocation in a time-varying channel environment cannot be effectively realized

Images