Optimization and robustness of structural engineering systems

This paper outlines several strands in the historical development of structural optimization techniques. The first is indirect via simple optimality conditions such as uniform density of stress or strain for achieving a technical objective like minimum weight, maximum stiffness or minimum strain energy. In the case of prestressed structures, zero net displacement of the concrete in flexure at service loads may be the criterion, and in the case of structural vibration damping, maximum dissipation of energy under dynamic loads. The second is direct via mathematical optimization of a more comprehensive, economic objective such as minimum cost or maximum utility under multiple loading conditions. Further development has led to automated and interactive design and optimization of nonconvex objective functions using improved optimization techniques. A more general, systems theory has been formulated with application to both strongly and weakly interacting systems, including structures, taking uncertainty into account. It relates overall utility of the system to its efficiency, diversity and uncertainty or robustness, and enables an overall optimum to be obtained for a given level of uncertainty and diversity. (C) 1997 Elsevier Science Ltd.