Abstract
In recent years, routing is considered one of the most challenging issues in MANET. The location of the stable node and the routing is based on the predicted locations that assists in establishing a routing path in MANET. The major intention of this paper is to detect the stable neighbor node in the MANET and also to establish stable multi-path routing for the various mobility patterns. Also, this paper deals with the data packet scheduling over multi-paths for balancing the load and forward the entire packets in the least broadcast time. The proposed approach elucidates four significant phases: stable node prediction, determination of stability measure, route exploration and packet dissemination. Initially, the stable node is predicted using the RMSG approach. Here, the stable neighbors are selected via Garson’s pruning based Recurrent neural network with a Modified seagull optimization algorithm (RMSG). In the route exploration phase, the path is created among the source and destination by the stable node. If any of the links fails, the route recovery process is established. Finally, the structure is formed for data packet distribution across the multipath. The proposed approach is evaluated by few performance measures such as throughput, packet delivery ratio, end-to-end delay, routing overhead energy consumption, and optimal path. This result describes that the proposed approach outperforms other state-of-art approaches.
Distributed Multipath Routing with Handling of Congestion in TCP/IP Networks using Neural Network
