Reliability-based optimization on a graph-theoretic modelling foundation

This paper provides a recipe for reliability-based design of steady-state systems to provide a minimum cost design subject to the constraints on the probability of designated failure events. The method is efficient and allows for arbitrary probability distributions. The method integrates graph-theoretic models and symbolic programming for model building with structural mechanics reliability methods for quickly approximating probabilities. The resulting models can be passed to any appropriate optimization routine. In this paper, the method is described along with a supporting example