Back Propagation Neural Network Driven African Vulture Routing Optimization in Software-Defined Networks

In large-scale networks, routing optimisation is critical for satisfying Quality-of-Service (QoS) requirements. Optimal routing is one of the most critical issues in Software-Defined Networks (SDNs), which determines the optimum method for moving data within the network in terms of performance, traffic loads, and energy. To increase the effectiveness of the routing in SDNs, the integration of Back Propagation Neural Network (BPNN) and African Vulture Routing Optimization (AVRO) algorithm is proposed in this research. Initially, the population is initialized and the Cluster Heads (CHs) are selected using AVRO algorithm so that the energy consumed during transmission is minimum and the distance between the transmitting and the receiving node and the Base Station (BS) is minimized. The BPNN algorithm estimates traffic density and adapts the possible routes of the traffic in real time. Experimental results show that the proposed BPNN-AVRO achieved Packet-Drop-Ratio (PDR) with 97.48%, End-to-End-Delay (EED) with 1.32ms, and throughput with 487 bits/s compared to existing methods Deep Q-Network and Traffic Prediction based Routing Optimization (DTPRO) and Improved ant colony optimization (IACO).