Enhanced Multipath Routing with Congestion Avoidance for 802.11E Based Mobile Adhoc Networks

Distributed Multipath Routing with Handling of Congestion in TCP/IP Networks using Neural Network

“…Results of testing are given in Tables (3), and (4), for computer networks (CN1 and CN2). Success rate of testing of the neural network (NN) for the computer networks (CN1 and CN2) is shown in Table (5).…”

“… P. Terdal, V. D. Mytri, Damodaram, and Uday. S. B [5], proposed a routing mechanism that can take advantage of the service differentiation offered by Enhanced Distributed Channel Access (EDCA) MAC and the same time overcome its limitation under heavy data. This work measures the available bandwidth of the high priority access categories, energy level and contention level experienced at the intermediate nodes to determine robust paths and divert the audio-video stream along such less congested paths, to ensure better end-to-end delay and throughput.…”

“…At first, the neural network determine all neighbor nodes, then the node chooses the best neighbor node to sent packet to the destination node, and due to any congestion or fault the node changed to the next best neighbor node, to solve the fault or congestion problem. For example, in the computer network (CN1), if the node 2 has packet to be sent to destination node 4, then the neural network (NN) at node 2 receives destination 4, delay time of link 2−1, 2−3, and 2−5, and rate of packet dropping for the neighbor nodes (1,3,5). It determines all neighbor nodes to reach node 4 quickly ordered depends on its precedence.…”

Distributed Multipath Routing with Handling of Congestion in TCP/IP Networks using Neural Network

“…Figure(5) error versus number of epochs of the feed forward neural network (NN) for the computer network (CN2).…”