5G Resource Allocation Enhancement Via Resnet-Inception-V2 With Non-Linear Analysis

Abstract

The building of 5G networks becomes quite challenging in resource allocation due to the increasing need for fast, low-latency connectivity. Effective use of resources determines both network performance and quality of service guarantee. This work solves resource allocation in 5G networks by combining Stochastic Paint Optimizer (SPO) with ResNet-Inception-V2 using non-linear analysis. Feature extraction and network traffic pattern categorization make use of ResNet-Inception-V2 architecture, which has capacity to record complicated traits and correlations. The Stochastic Paint Optimizer is applied to maximize resource allocation by means of an iterative, stochastic technique balancing exploration and exploitation, hence addressing the allocation problem. Non-linear analysis provides model and prediction of the non-linear interactions between network properties and resource constraints. The proposed method is tested in a simulated environment including actual 5G network traffic. Results indicate considerable improvement in network performance and resource economy. More exactly, the method improves resource allocation efficiency by 15% over traditional methods. By 12%, latency is reduced; by 18%, general network capacity increases. These results show how well contemporary deep learning architectures and optimization strategies blend to maximize 5G resource allocation.