Optimization Techniques for Damage Detection

This article is concerned with techniques for optimization problems in the context of structural health monitoring. Many techniques are available for solving various classes of problems, including the computationally demanding combinatorial optimization problems which arise in principled approaches to feature selection or sensor optimization. This article briefly describes the classical, gradient-based approaches that are well represented in the mathematical literature, and then goes on to discuss, in much more detail, the heuristic methods inspired by physics and biology, which have (comparatively) recently proved to be very powerful for combinatorial problems.

Keywords: optimization; combinatorial optimization; steepest descent; simulated annealing; genetic algorithms; differential evolution; ant colony metaphors; immune system metaphors; sensor placement; damage identification