Evolutionary search and convertible agents for the simultaneous type and dimensional synthesis of planar mechanisms

John C. Oliva and Erik D. Goodman
Proceedings of the 11th Annual Conference on Genetic and Evolutionary Computation, (Montreal, Québec, Canada, July 08 - 12, 2009). GECCO '09. ACM, New York, NY, pp. 1577-1584.

In the field of mechanical engineering, synthesizing a mechanism to perform an intended task is deceptively complex. In this paper, a novel approach to automated mechanism synthesis is described which uses an evolutionary search algorithm and a technique called "convertible agents" to simultaneously find the most appropriate mechanism type for a given problem, while finding an optimum set of dimensions for that mechanism to complete a specified task. The search was limited to four-bar, Stephenson, and Watt types of planar, single-degree-of-freedom mechanisms, although the method is readily scalable to include any number of different types. Several case studies are described which illustrate the effectiveness of the method. The developed convertible agent approach is well suited for evolutionary design applications in which there are a small number of distinct topological possibilities each with parametric variables to be optimized.
 Read Abstract