Worst-Case Tolerance Analysis in Static Field Problems

This paper introduces an evolutionary approach to worst case tolerance analysis of static field problems ruled by the Poisson equation, over suitable spatial domains and with given boundary conditions. The approach ensures a better estimation of the true worst case with respect to classical stochastic methods, especially if parameters are affected by large uncertainties and whenever the performance function of interest is nonlinear or, even worse, nonmonotonical with respect to parameters.