The motive of current study is to design a novel stochastic radial basis deep neural network for the food chain (FC) supply system with Allee impacts. Two hidden layers are taken and the radial basis is performed in both hidden layers. The FC supply model is divided into three species, top-predator, predator, and the population of prey. Three different model’s cases are numerically presented, and the Bayesian regularization (BR) is used as an optimization. The novel process is designed by using the optimization based on the BR, radial basis along with twenty-five and forty-five neurons in first and second hidden layer. The dataset is obtained by the explicit Runge-Kutta scheme, which is used to decrease the mean square error by selecting the data as 68% for training, 20% for testing and 12% for endorsement. The results overlapping, small absolute error and statistical measures are used to perform the correctness of the scheme for solving the FC supply system.
Sabir et al. (Thu,) studied this question.