Evolutionary backtracking search optimization for the dynamics of nonlinear benchmark physical systems

Abstract

ABSTRACT A novel application of bio-inspired computing heuristics with a backtracking search optimization algorithm (BSA) is presented for solving a system of nonlinear equations arising in a benchmark physical environment representing an interval arithmetic benchmark model, a kinematic application model, a neuro-physiology problem, combustion model and a chemical equilibrium system. The approximation in a mean square sense is exploited to formulate the residual error-based merit function for the systems of nonlinear equations. The design parameters of the merit function are calculated by using the strength of the optimization efficacy of BSA. Five variants of BSA are implemented to find the solution of benchmark nonlinear physical systems and comparative studies with state-of-the-art counterparts certify the worth of BSA in terms of precision, convergence, stability and complexity measures on single and multiple autonomous executions.