Swarm-Intelligence Optimization Method for Dynamic Optimization Problem

Mathematics

Yuanbin Mo¹  Haidong Guo²  Yucheng Lyu²  Yanyue Lu³  Rui Liu⁴  Yuedong Zhang⁴ 

[1] Guangxi Key Laboratory of Hybrid Computation and IC Design Analysis, Guangxi Minzu University, Nanning 530006, China;Institute of Artificial Intelligence, Guangxi Minzu University, Nanning 530006, China;School of Chemistry and Chemical Engineering, Guangxi Minzu University, Nanning 530006, China;School of Electronic Information, Guangxi Minzu University, Nanning 530006, China;

关键词: dynamic optimization;    swarm intelligence;    control variable parameterization;    nonlinear programming problem;    sparrow search algorithm;   

DOI  :  10.3390/math10111803

来源: DOAJ

【 摘 要 】

In recent years, the vigorous rise in computational intelligence has opened up new research ideas for solving chemical dynamic optimization problems, making the application of swarm-intelligence optimization techniques more and more widespread. However, the potential for algorithms with different performances still needs to be further investigated in this context. On this premise, this paper puts forward a universal swarm-intelligence dynamic optimization framework, which transforms the infinite-dimensional dynamic optimization problem into the finite-dimensional nonlinear programming problem through control variable parameterization. In order to improve the efficiency and accuracy of dynamic optimization, an improved version of the multi-strategy enhanced sparrow search algorithm is proposed from the application side, including good-point set initialization, hybrid algorithm strategy, Lévy flight mechanism, and Student’s t-distribution model. The resulting augmented algorithm is theoretically tested on ten benchmark functions, and compared with the whale optimization algorithm, marine predators algorithm, harris hawks optimization, social group optimization, and the basic sparrow search algorithm, statistical results verify that the improved algorithm has advantages in most tests. Finally, the six algorithms are further applied to three typical dynamic optimization problems under a universal swarm-intelligence dynamic optimization framework. The proposed algorithm achieves optimal results and has higher accuracy than methods in other references.

【 授权许可】

Unknown