Models are being developed to account for the conditional reliability of structures. For example, successful past performance of a nondeteriorating structure has a proof load effect that increases its reliability estimate for subsequent service and decreases the likelihood of gross error in strength. For deteriorating systems, uncertainty of the state of the system grows significantly with time. Thus, reliability models are being developed that can combine new information, such as test or inspection results, with older estimates of serviceability, and enable the costs and expected value of new test information to be balanced against each other, and against nontest options.

For economical reasons, many structural members are produced in fixed sizes rather than in a continuous supply of structural shapes. This presents an interesting optimization problem, namely, how to design an assortment of profiles or sizes to best satisfy a structural demand. One solution approach is to minimize the material waste from overdesign that occurs when standardized sizes are selected rather than ``made-to-order'' cross sections. Related problems include the modeling of economy of scale and the optimal consolidation of production materials.

The ability of a structural control system to maintain stability is measured by its robustness. When parameter uncertainty is modeled by a probability distribution, robustness becomes analogous to system reliability, in which the joint distribution is of dimension equal to the number of uncertain parameters and the failure boundary is defined by the onset of instability. In the approach taken, root loci provide a mapping of system performance into the parameter space, defining failure boundaries. System reliability methods are then used to assess system robustness from the set of all modal failure regions. A further benefit is the ability to revise the robustness measure when new information from testing or observation is obtained.

Reliability models of structural design are extended to cover the use of load testing for test-based design of new structures or for the evaluation of existing structures. Approaches are being developed to incorporate into de cision making the information from sample tests, single-mode or multimode proof tests, and other sources. These and other load-testing models allow both test and nontest information on strength and safety to be used in the evaluation of a structure. Procedures for design by testing can be assessed for their likely effects on achieved reliability, and design resistance factors can be found that account for test uncertainties.