Model Order Reduction of MIMO Linear System Using Grey Wolf Optimization

Scientists and engineer are being encountered regularly with the workout, analysis and amalgam of real-life complications. The component of an actual framework can be broken down by the arrangement of differential conditions from which complex model of high order can be obtained in the state space structure or in transfer function form which are called frequency domain representation or time domain, respectively. Difficulties may be posed by a mathematical model of high order in its amalgam, analysis and recognition. Thus, a demand occurs for an organized method to develop a reduced model which preserves the main subjective properties of high-order framework like steadiness, dependability, good time and frequency response matching as in the original framework. Model reduction helps in simplifying the system, reduces computational efforts and makes the design of the controller numerically more efficient. This work is encouraged by nature-based algorithm for order reduction by grey wolf optimization. This algorithm mimes the ranking-based leadership and hunting conduct of grey wolves. The ranking-based leadership is simulated by employing four types of grey wolves such as alpha, beta, delta and omega. Grey wolf optimization technique is used for determining the denominator polynomial of the reduced model, while Pade approximation technique is used for finding the numerator polynomial. The approach is proposed for both single input single output (SISO) and multi-input multi-output (MIMO) systems. This technique assures the stable reduced models of stable systems. The technique is protracted to the issue of reducing a high-order multivariable system defined by its matrix transfer function. Some examples from the literature are taken for illustrating the feasibility of the method. ISE has been taken as target work for looking at the performance of the proposed strategy with other existing techniques in the writing.