GA based decomposition of large scale distributed model predictive control systems. (December 2016)

• Record Type: Journal Article
• Authors: Xie, Lei; Cai, Xing; Chen, Junghui; Su, Hongye
• Is Part Of: Control engineering practice. Volume 57(2016)
• Page Start: 111
• Page End: 125
• Publication Date: 2016-12
• Subjects: Distributed model predictive control -- Genetic algorithm -- Input clustering decomposition -- Input–output pairing decomposition -- System decomposition; Automatic control -- Periodicals; 629.89
• DOI: 10.1016/j.conengprac.2016.08.016
• Languages: English
• ISSNs: 0967-0661

Abstract

Abstract: A novel method is proposed to find the optimal decomposition structure of distributed model predictive control (DMPC) systems. The input clustering decomposition (ICD) is first developed to minimize the coupling effects of subsystems and average the computational balance of each subsystem. To select the inputs and outputs in each subsystem, the input–output pairing decomposition (IOPD) is done. Then the genetic algorithm is used to solve decomposition problems for ICD and IOPD. The proposed method can achieve efficient coordination. Its structure is more flexible than the traditional DMPC. Two examples are used to show the abilities of the proposed method.