Improved Harris hawks optimization algorithm based on random unscented sigma point mutation strategy

Wenyan Guo

doi:10.1016/j.asoc.2021.108012

2021-12-01

PRE

DOI：10.1016/j.asoc.2021.108012

原文链接

原文求助

摘要

En 中文

The Harris hawks optimization algorithm (HHO) is a new swarm intelligence algorithm for simulating the capture and attack of the hawks which has good global exploration and local exploitation capabilities. In order to further improve the optimization performance of the algorithm, quasi-opposite learning and quasi-reflection learning strategies performed according to probability are involved in the attack phase to enhance the diversity of the population and accelerate the convergence rate of HHO while a logarithmic nonlinear convergence factor is designed to balance the ability of global search and local optimization of the algorithm. Furthermore, in order to avoid the algorithm falling into a local optimum, using the characteristics of the unscented transform (UT) to estimate the mean and variance of a random variable function can achieve second-order accuracy, a new strategy for generating random symmetric sigma points is designed to mutate the current best individual in the visible range, at last, an improved Harris hawk algorithm (IHHO) based on random unscented sigma point mutation is proposed. The new stochastic UT ensures the random exploitation of the algorithm and has a certain theoretical support, which overcomes the theoretical deficiency of the stochastic optimization algorithm to some extent. The numerical optimization ability of IHHO is verified by CEC2017 benchmark functions, CEC2020 benchmark function and fifteen standard test functions. Finally, the practicality and effectiveness of the IHHO algorithm are verified by three engineering constraint optimization and the discounted {0-1} knapsack problem. (c) 2021 Elsevier B.V. All rights reserved.

Keyword:

Harris hawks optimization algorithm

Quasi-opposite learning

Quasi-reflection learning

Unscented transformation

Discounted {0-1} knapsack problem