Neuro-swarm intelligence to study mosquito dispersal system in a heterogeneous atmosphere

The current work is to design a novel neuro-swarming computational approach for numerical treatment of mosquito dispersal system based on the heterogeneous environment by functioning the approximation capability of artificial neural networks (ANNs), global search exploitation of particle swarm optimization (PSO) together with speedy local search with interior-point (IP), i.e., ANN-PSOIP approach. In the designed ANN-PSOIP approach, the construction of objective/merit function is conducted by defining a mean square error function with continuous mapping in terms of differential models of ANNs for governing ODEs of mosquito dispersal systems and these trained nets are proficient initially with PSO and final with IP approach. The precision, correctness, stability and reliability of the proposed ANN-PSOIP approach is accredited via comparative trainings with the Runge–Kutta scheme. The obtained valuations further authenticate the accuracy as well as convergence of the designed ANN-PSOIP.