A Dual Approach to Optimum Truss Design

Abstract This paper considers the problem of finding member sizes which minimize the weight of a pin-jointed truss of fixed geometry while satisfying constraints upon joint displacements, member stresses, and minimum sizes. Aspects of both mathematical programming methods and optimality criteria methods for designing large trusses are discussed. The optimality criteria approach is further extended and the whole truss design problem is recast into a new dual formulation in which constraint activity levels are used as variables in a mathematical programming solution method. This new dual formulation unifies both the optimality criteria and mathematical programming approaches to the problem of truss design. The paper is theoretical in nature, being largely devoted to a proof of the dual method. A discussion of the likely implications and usefulness of the dual approach to truss design is given with comments upon its possible modes of practical use.