Abstract:
The purpose of this article is twofold: (1) to apply an ant-gradient (AG) algorithm, which is a combination of Ant Colony Optimization and gradient search to the mechanism synthesis problem and (2) to introduce the idea of hybrid task mechanisms. A brief overview of the AG algorithm is presented and applied to dimensional synthesis of planar four-bar mechanisms. The algorithm is tested by way of two benchmark examples. Compared with the corresponding results generated by other optimization schemes, the AG search rendered the best solution of all. The concept of hybrid task synthesis and a scheme to determine the crank angles during a motion generation phase and prevent order defects is then introduced and demonstrated by way of a practical example solved by the AG algorithm. © 2006 Elsevier Ltd. All rights reserved.