Simultaneous Type and Dimensional Synthesis of Mechanisms by Genetic Algorithms

Abstract

Abstract Based on global search/optimization techniques called genetic algorithms, a new approach to the conceptual design of mechanisms is proposed. Emulating “natural” genetics and selection, competing types of mechanisms are treated as species and different dimensions are evaluated for each type simultaneously. The problem formulation is straight-forward, similar to an optimization problem. The result may be a spectrum of similar or different mechanisms, depending on whether multiple optimal solutions exist. The computational power required to utilize this approach is acceptable for most conceptual design problems.