Research On Optimization Strategy Of Distributed Model Predictive Control System Based On GA-PSO

Model predictive control through rolling optimization,online forecasting makes it an advanced control algorithm that adapts to the characteristics of complex industrial processes.However,with the advancement of science and technology,in the actual industrial production,due to the requirements of computational burden and coordination and the geographical dispersion of the control system,it is very difficult to control in a centralized manner,and it is prone to failure.In this context,Distributed Model Predictive Control(DMPC)with distributed structure and model predictive control features has become an important means to deal with complex industrial control problems.Distributed model predictive control can reduce the computational burden of subsystems and enhance the scalability of the system.At present,DMPC is still in the early stage of research,and there are still many problems to be solved,such as achieving better control effects under the possible less communication burden;the decomposition of DMPC system and the coupling relationship between processing subsystems;how to design the weight matrix of the system to achieve better control results.Based on the research of distributed model predictive control,this paper optimizes and improves the control performance of the system.The specific work is as follows:1.In order to decompose the distributed model predictive control system,a hybrid optimization algorithm using genetic algorithm and particle swarm optimization algorithm is proposed to decompose the system.The hybrid optimization algorithm combines the advantages of both algorithms.Compared with the single optimization algorithm,the convergence speed is faster and the precision is higher.The input grouping problem is essentially an optimization problem.Grouping the input of the control system can effectively eliminate the coupling between the system inputs and balance the communication burden between the subsystems.Therefore,the input grouping of the control system can complete the structural decomposition of the system.This paper defines a system disassembly objective function to solve the input grouping problem,and verify the effectiveness of the method by grouping experiments on some inputs in the hot dip galvanizing line control system.2.In order to optimize the control weight of the system,a method based on genetic particle swarm optimization is proposed.The mutation and crossover operation of genetic algorithm are introduced into the particle swarm optimization algorithm to obtain a better genetic particle swarm optimization algorithm.Taking the multi-agent system as the research object,the control weight of the iterative distributed model predictive control system is optimized.This paper defines an optimization objective function to optimize the control weight of the system,and simulates the power system through a ten-area power system to verify the effectiveness of the method.