Adaptive multi-path routing based on an improved leapfrog algorithm

Congestion in wireless sensor networks (WSNs) can result in the phenomenon of packet loss, which in turn reduces throughput and wastes energy. Therefore, congestion in WSNs needs to be controlled to achieve the goals of high energy efficiency, prolonged system lifetime, and better fairness and quality of service. We propose an adaptive multi-path approach based on an improved leapfrog algorithm to solve the transmission-congestion problem in WSNs. Specifically, this paper establishes a path-satisfaction model that considers the predicted degree of congestion, the remaining energy, and the minimum number of hops. The algorithm is updated discretely during the local optimization process, and a variable learning factor is introduced. The memetic information of individual frogs is optimized with a threshold-selection strategy that encourages leaping from weaker individuals to better ones. In global optimization, the search orientation of each memeplex is based on exchanging and recombining information with other memeplexes, and a multi-path routing idea is introduced to select the optimal path. Simulation results show that this method performs well in real time, significantly improves energy efficiency, and prolongs the network lifetime.