Analysis of Congestion Control Techniques for Improving QoS using Machine Learning Approach in MANET

Congestion is the main reason why connections fail when there are too many people on the network. This leads to the loss of nodes and modify the way setup of Mobile Ad-hoc Network (MANET). As we know, MANET is a network with no permanent infrastructure. It is made up of a temporary network that doesn't need a central manager or the usual support equipment found in traditional networks. To assure the longevity of MANET, networks can be constructed anywhere. Buffer overflow and congested links are caused by severe network demand, resulting in significant packet loss. Proper routing results in packet delays and has an effect on the MANET protocol's packet delivery rate. Machine learning (ML) is a part of Artificial Intelligence (AI) which is able to improve the QoS of routing in networks. In this paper, we discuss the latest research work to learn some innovative concepts and focus on doing something new in the field of Machine Learning in MANET. A well-balanced strategy is a method for avoiding network congestion. Load balancing allows network consumption to be optimized, packet latency to be reduced, and packet distribution ratio to be enhanced. Shifting loads from congested to less congested paths enhance the network's overall efficiency. MANET routing protocols prefer paths with less hops and exclude paths with more hops, but in extreme load conditions, they prefer paths with less hops. Furthermore, we investigate congestion management routing options in order to suggest a new approach in the future for identifying and preventing congestion risk in MANET.