PSO and DE Based Model Order Reduction and Discrete Time Structure Specified Controller Design

Background/Objectives: The objective is to reduce the order of the plant and design discrete time PID controller using Particle Swarm Optimization (PSO) and Differential Evolution. Methods/Statistical analysis: Two popular optimization techniques – PSO and Differential Evolution (DE) are used to obtain the reduced order model of the physical plant and its discrete time structure specified PID controller. The objective function used for model order reduction is an integral square error between outputs of original and reduced order model. Whereas, the controller designs, integral square error between output and reference signal is chosen for tracking purpose. Findings: Nominal models for physical systems are highly complex and their order is usually very high. Designing controllers for such models in industries require huge amount of cost and large number of hardware components. The optimization techniques will prevent this issue, as shown in this paper. The optimal reduced model and optimal controllers are extremely useful for the industrial applications. Application/Improvements: In this paper, this methodology is proposed to design discrete time controller for a process of eight orders. But, this approach can be effectively applied to the complex plants for implementation of controller.