Affiliations: [a] School of Artificial Intelligence, Guangxi Minzu University, Nanning, China
| [b] Guangxi Key Laboratory of Hybrid Computation and IC Design Analysis, Guangxi Minzu University, Nanning, China
| [c] Guizhou Police College, Guiyang, China
| [d] School of Computer Science and Technology, Aba Teachers College, Wenchuan, China
Abstract: Dynamic optimization problems exist widely in chemical industry, and its operational variables change with the evolution of both space and time. Therefore, dynamic optimization problems have important research significance and challenges. To solve this problem, a multi-strategy mayfly optimization algorithm (MMOA) combined with control variable parameterization method(CVP) is proposed in this paper. MMOA introduces three improvements on the basis of the original algorithm, namely, circle chaos crossover strategy, center wandering strategy and boundary correction strategy. The hybrid strategy can better balance the exploration and exploitation ability of the algorithm. Based on MATLAB simulation environment, MMOA was evaluated. The experimental results show that MMOA has excellent performance in solving precision, convergence speed and stability for the benchmark function. For the six classical chemical dynamic optimization problems, MMOA obtained the performance indexes of 0.61071, 0.4776, 0.57486, 0.73768, 0.11861 and 0.13307, respectively. Compared with the data in the previous literature, MMOA can obtain more accurate control trajectory and better performance indicators. It provides an effective way to solve the dynamic optimization problem.
Keywords: Chemical dynamic system, process control, dynamic optimization, mayfly optimization algorithm, control vector parameterization
