There is a new book out on Robustness Analysis in Decision Aiding, Optimization, and Analytics, edited by Michael Doumpos, Constantin Zopounidis, and Evangelos Grigoroudis.

André and me participated with a chapter on Performance Analysis in Robust Optimization. The basic idea can be encountered, even though implicitly, in many papers on robust models: Now that we have found a solution we consider robust, how can we evaluate how good it performs?

One way is to calculate how much the robust solution costs in the nominal scenario, compared to how much the best possible solution in the nominal scenario costs. But that is clearly one-sided: Of course the robust solution will costs more, but it is much more interesting to see how much we will *save* by using the robust solution.

There is not one single, fair method to evaluate a solution. We present different approaches to do so, and discuss advantages and disadvantages. As a side node, the average-case solution often performs not quite bad at all!

The book contains many more papers that will be highly interesting to the community. To highlight two: There is a survey on robust combinatorial optimisation which is exactly what was needed at the moment, and another survey on The State of Robust Optimization.