A neural network-based load balancing optimization method for ultra-dense heterogeneous networks

Abstract

The invention provides an ultra-dense heterogeneous network load balance optimization method based on a neural network; a low-complexity ultra-dense heterogeneous network downlink user connection method combines an ART2 type neural network and a cost-based distributed method; a cost offset value of all small stations is adjusted in a combined mode, and the problem of load balance in the ultra-dense heterogeneous network is solved. According to the method, an initial value is set by adopting a classification of the ART2 type neural network, so that the number of iterations and calculation complexity can be greatly reduced, and the throughput rate of users at the edge of the base station and the middle of the base station can be improved; the load of the base station between the cross layerand the same layer is automatically balanced, the number of iterations of the load balancing iteration method is further remarkably reduced, and the method is more suitable for a fast, complex and changeable practical situation.

Description

technical field [0001] The invention belongs to the technical field of wireless networks in mobile communications, and relates to a load balancing optimization method, in particular to an ultra-dense heterogeneous network load balancing optimization based on adaptive resonance theory 2 (Adaptive Resonance Theory, ART2) neural network in a wireless communication system method. Background technique [0002] An ultra-dense heterogeneous network with densely deployed low-power small cells at the same frequency within the coverage area of ​​macro cells is an effective method to improve the spectrum utilization and network capacity of the fifth generation mobile communication (5G) network. In the commonly used serving cell selection criterion—the maximum power receiving criterion, each user selects the cell with the strongest received signal power as the serving cell. However, in a heterogeneous network, the power difference between the large station and the small station is rela...

AI Technical Summary

Problems solved by technology

By relaxing the problem, a low-complexity cost-based distributed method can be obtained to converge to an approximate optimal solution. However, the convergence speed of this cost-based distributed user connection method depends on the selection of iteration parameters. For the complex situation of the actual network, it is impossible to adjust the iteration parameters for real-time load changes, and the convergence speed cannot be guaranteed

Images