A multi-objective design tree approach for the optimization of mechanisms

Abstract A design tree structure is proposed for the optimization of mechanisms which considers both geometric and topological change. The optimization is performed by a nonlinear goal programming algorithm which allows for the consideration of multiple, conflicting design criteria. The combination of the tree structure and nonlinear goal programming provides a unique design environment which can handle issues such as uncertainty in design parameter values, fuzziness in the design criteria or constraints and discrete or integer design variables. The method enable design for latitude to be investigated at the conceptual design phase. The objective is to move away from the rigid structure imposed by conventional nonlinear programming and to allow “real” design issues to be addressed. The technique is applied to the design of a function generating mechanism with uncertainty in both exact coupler point location and link lengths. A four bar mechanism, an inverted slider crank mechanism and a general six-bar mechanism are considered as design alternatives.