A novel gray wolf optimizer with RNA crossover operation for tackling the non-parametric modeling problem of FCC process

This work proposes a novel gray wolf optimizer (GWO) with RNA crossover-operation and adaptive control parameter scheme (named as RNA-GWO). The RNA crossover-operation can better enhance the population diversity of RNA-GWO, it is designed according to the special pseudoknot structure of RNA molecule. The adaptive control parameter scheme is proposed to replace the linear one to balance the exploration and exploitation capacities during the optimization process of RNA-GWO. The effectiveness of RNA-GWO is verified by a suite of IEEE CEC 2017 benchmark functions. The experimental results prove that RNA-GWO can obtain higher accuracy solutions than the other six meta-heuristic algorithms. The RNA-GWO is also employed to solve the parameter problem of the wavelet neural network (WNN) non-parametric modeling method and applied to model the FCC process. The simulation results demonstrate that the RNA-GWO can provide favorable parameters for WNN, the model outputs of RNA-GWO optimized WNN can better agree with the experimental data of FCC process. (C) 2021 Elsevier B.V. All rights reserved.

Automated summary

This work introduces a novel gray wolf optimizer RNA-GWO with RNA crossover-operation and adaptive control parameter scheme. The effectiveness of RNA-GWO is verified through experimental results and it is successfully applied to solve the parameter problem of the wavelet neural network non-parametric modeling method.