Behavior of fully stressed design of structures and its relationshipto minimum-weight design.

The iterative method of analysis and fully stressed redesign does not always converge to a minimum-weight design. Using the Kuhn-Tucker optimality condition of nonlinear programing, the necessary condition for the equivalence of the two methods of design is found and a computationally practical method for verification of the optimality of fully stressed design is obtained. Where the fully stressed design is not optimum a method is suggested for determination of the optimum which reduces the dimensionality of the problem by dividing the design variables into free and fully stressed ones. Then, the optimization is decentralized into an optimal search on free design variables and fully stressed design of the rest of the variables.