Random Early Detection Protocol for Congestion Prediction and Avoidance in MANET under TCP

Transmission Control Protocol (TCP) ensures data transfer reliability for all internet end-to-end data stream services. The information will be transferred throughout the Mobile Ad Hoc Network (MANET). Data transmission becomes unusual when there is congestion in the path of the data while being transferred. If congestion occurs packet loss got occurred and precision will also be in the slanting state. To solve this issues, Artificial Neural Network (ANN) through Random Early Detection Protocol (RED) is developed for prediction also for avoiding the congestion flow in the path of data transmission. Initially the data’s will be transferred between the networks from the source to the destination. As a data is initiated from the transmission side and the other side of the receiver will be waiting for the acknowledgement. As soon as the acknowledgement got received, the function of congestion prediction will be initiated. This congestion prediction will be held by the ANN for transforming data transmission into an efficient data transmission. Furtherly, Golden Jackal Optimization (GJO) is used for selecting the optimal weights present in the ANN classifier. After the identification of the congestion is existed then it will be avoided using the REDP. The simulation analysis shows the suggested method gives improved performance of TCP in MANET.