Critical consideration of the suitability of randomized optimization methods: Power system topology estimation problem

Many engineering areas are benefiting from the abundance of powerful computing capabilities, which allows for the application of computationally intensive techniques, such as computational intelligence. Such techniques are frequently universal and do not require excessive tailoring for the specific problem. That makes it attractive for wide use, but is however, sometimes, unjustified. This paper discusses the necessity to evaluate whether a particular problem is suitable to be addressed using such techniques, for the particular case of combined power system state and topology identification.