Intelligent Network Optimisation for Beyond 5G Networks Considering Packet Drop Rate

To meet the growing expectations for fast and dependable connectivity, Novel approaches such as reinforcement learning-based resource allocation and network slicing are essential to consider. Enhancing network intelligence through the use of deep reinforcement learning and machine learning may increase capacity, lower latency and congestion, improve energy efficiency, and open up new revenue opportunities and business models. However, the current body of research falls short in terms of thoroughly exploring network slicing datasets and taking packet drop likelihood into account when allocating resources. With reference to current benchmarks, this paper presents a network slicing strategy and illustrates its efficacy using seven machine learning algorithms (ANN, SVM, RF, DT, GNG, LDA, and RT). To prioritize packets that run the risk of being dropped, a priority algorithm is also developed. To improve network performance, a resource allocation technique is used that is based on the mathematical study of packet drop rate. Utilizing deep reinforcement learning and genetic algorithms, the distribution of tasks across Cloud and Edge resources is further improved according to network slice characteristics. In spite of high network traffic, this guarantees constant service availability.