“…PNLBSS methods can rapidly converge based on standard gradient descent methods; however, satisfactory results are sometimes difficult to be obtained due to the existence of many local minima in complex nonlinear spaces. Over the last decades, evolutionary algorithms (EAs) have gained considerable attention to obtain a global optimal solution, such as particle swarm optimization (PSO) (Yang et al., 2016), memetic algorithm (Mousavi and Alfi, 2015), genetic algorithm (Deb and Srivastava, 2012), harmony search algorithm (Ameli et al., 2016), water cycle algorithm (Pahnehkolaei et al., 2017), and search group algorithm (Noorbin and Alfi, 2018), which have been successfully applied to solve optimization problems in practical applications. In this study, PSO (Pornsing, 2014) is used to update the parameters in PNLBSS, and crossover and mutation are used in PSO to avoid particle premature convergence.…”