A deep learning resource allocation scheme for physical layers in communication systems

Abstract

Background/Objectives: In this paper, we have studied the deep learning channel allocation scheme based on user - patterned big data. It is an adaptive channel capacity management scheme through analysis of user usage pattern, which can allocate the optimal channel compared to existing system. We show that the proposed method can effectively manage users and improve the performance of the whole system. Methods/Statistical analysis: In a typical environment, channels are allocated considering user traffic. However, when a large number of users are gathered, allocation based on the channel capacity is mainly performed without considering the usage pattern of each user. In such a case, traffic usage increases rapidly depending on the usage pattern of the user, causing traffic exceeding the channel capacity, resulting in a problem that the performance of the entire system is greatly degraded. As a solution to this problem, we proposed a deep learning resource allocation scheme that allocates different channels to usage patterns based on user usage pattern big data. Findings: It is confirmed that the network performance degradation due to channel interference does not occur much by allocating a relatively free channel in the channel interference based on the user group information. Also, the proposed system showed relatively uniform network performance compared to the existing system. Improvements/Applications: The proposed system is applicable to various networks. The number of users per network is rapidly increasing due to the increase of IoT devices, and it is time to manage network resources due to the introduction of many IT convergence technologies. In this environment, we can provide smooth service through the proposed method.