| Summary: | Due to massive data traffic, wireless network congestion is inevitable, causing network performance deterioration and data transmission quality reduction. By adjustingthe traffic flow at the source node along the optimal path, Wireless sensor networks strive to evade and overcome congestion. Recently, numerous evolutionary approaches for congestion detection and avoidance have been presented, although they do not improve WSN performance. A hybrid evolutionary approach is attempted to avoid congestion in WSN and improve network performance. We present an optimum cluster-based congestion aware multipath routing protocol (OCC-MP). An improved atom search optimization (IASO) approach for efficient clustering is introduced initially in OCC-MP. Then the HSIPO method is used for decision making, which computes each node's trust degree. The HSIPO algorithm selects the cluster head (CH) from a group of nodes. Then a deep recurrent neural network (DRNN) is used to monitor congestion and provide congestion aware routing. Finally, multiple simulation situations like various node densities andat various simulation rounds supported our proposed OCC-MP techniqueoutperformedcurrent methodologies with respect to energy consumption, throughput, traffic load overflow, delivery ratio, and number of nodes alive.
|
|---|