Multi-objective Optimization of a Steering Linkage Using Alternative Objective Functions

Abstract

This paper proposes comparing objective functions in multi-objective optimization of a rack-and-pinion steering linkage in which the optimum results can be affected by the types of objective functions especially when using evolutionary algorithms. The optimization of a steering linkage in the past studied minimizing a steering error and/or a turning radius which can be formulated as a single or multi-objective optimization problem. Steering error usually defines as the different angle between actual angle of steering wheels and the theoretical angle of the wheels according to the Ackerman’s principal. Alternatively, the steering error can be rearranged in form of a deviation of instantaneous center based on the same principal, but it still needs to clarify the advantage of using different steering error measures. As a result, it is our attention to study the effect of objective functions to the optimum results on multi-objective optimization of a rack-and-pinion steering linkage. The objective functions are assigned to simultaneously minimize a steering error (dimensionless angle or length) and a turning radius. The design variables are linkage dimensions. The design problem is solved by improving the hybridization of real-code population-based incremental learning and differential evolution (RPBIL-DE). The comparison shows that the alternative objective function can compare with the traditional objective, which leads to effective design of rack-and-pinion steering linkages.