An Evolutionary Approach to Solve a Novel Mechatronic Multiobjective Optimization Problem

Abstract

Summary. In this chapter, we present an evolutionary approach to solve a novel mechatronic design problem of a pinion-rack continuously variable transmission (CVT). This problem is stated as a multiobjective optimization problem, because we concurrently optimize the mechanical structure and the controller performance, in order to produce mechanical, electronic and control flexibility for the designed system. The problem is solved first with a mathematical programming technique called the goal attainment method. Based on some shortcomings found, we propose a differential evolution-based approach to solve the aforementioned problem. The performance of both approaches (the goal attainment and the modified DE) are compared and discussed, based on quality, robustness, computational time and implementation complexity. We also highlight the interpretation of the solutions obtained in the context of the application.