Multidimensional scaling analysis of generalized mean discrete-time fractional order controllers

The dynamics of discrete-time fractional order control systems depends on the method used for implementing the fractional derivatives and integrals. A reliable numerical approach adopts the generalized mean of the continuous to discrete conversion. The extra freedom provided by the proposed method must be carefully optimized by the user. This paper investigates the use of multidimensional scaling for evaluating and visualizing the performance of generalized mean discrete-time fractional controllers. Two alternative performance indices are adopted for comparing the time and frequency responses of the controlled system when adopting different combinations of the parameters. Numerical experiments with a fractional PID and two linear plants demonstrate the feasibility of the method for comparing and visualizing multiple test cases. (C) 2020 Elsevier B.V. All rights reserved.

Automated summary

This paper explores the performance evaluation and visualization of generalized mean discrete-time fractional controllers using multidimensional scaling, comparing time and frequency responses of controlled systems with different parameter combinations. Numerical experiments with fractional PID and two linear plants demonstrate the feasibility of this method for comparing and visualizing multiple test cases.