Design of PI controllers based on non-convex optimization (Q1298263)

The authors formulate the design problem as an optimization problem with a constraint on the maximum sensitivity and they exploit the structure of this optimization problem. They give iterative methods to solve the problem (if a solution exists, otherwise the methods indicate when there is no controller) and finally calculate explicitly 8 examples with very different transfer functions. They compare their method with existing ones and give hints where the ideas of the new method came from. It is quite interesting, that already in 1950 \textit{P. Hazebroek} and \textit{B. L. van der Waerden} [Trans. A.S.M.E. 72, 309-322 (1950)] discussed the use of optimization to solve special control problems of the kind handled in this paper.