Swarming Computational Procedures for the Coronavirus-Based Mathematical SEIR-NDC Model

The motive of the current work is related to solving the coronavirus-based mathematical system of susceptible (S), exposed (E), infected (I), recovered (R), overall population (N), civic observation (D), and cumulative performance (C), called as SEIR-NDC. The numerical solutions of the SEIR-NDC model are presented by using the computational framework of artificial neural networks (ANNs) together with the swarming optimization procedures aided with the sequential quadratic programming. The swarming procedure based on the particle swarm optimization (PSO) works as a global search, while the sequential quadratic programming (SQP) is used as a local search algorithm. A merit function is constructed by using the nonlinear dynamics of the SEIR-NDC mathematical system based on its 7 classes, and the optimization of the merit function is performed through the PSOSQP. The numerical expressions of system are accessible with the ANNs using the PSOSQP optimization with 30 variables. The correctness of the stochastic computing scheme performances is verified by using the comparison of the obtained performances of the mathematical SEIR-NDC system and the reference Runge–Kutta scheme. Furthermore, the graphical illustrations of the performance indices, absolute error, and convergence curves are derived to validate the robustness of the proposed ANN-PSOSQP approach for the mathematical SEIR-NDC system.