Iterative optimization RCO: A "Ruler & Compass" deterministic method
Maurice Clerc · HAL (Le Centre pour la Communication Scientifique Directe) · 2024
We present here the basic version of an iterative optimization algorithm that is deterministic, needsjust one parameter, and often capable of finding a good solution after very few evaluations of the fitnessfunction. We explain its principles using a multimodal one-dimensional problem. For such problems, it couldbe applied using nothing more than a ruler and a compass. We provide classical examples illustrating thealgorithm’s properties, as well examples where it fails. As this version does not address possible stagnation,it is really only useful for low-dimensional problems (typically at most ten), where each evaluation of aposition in the search space is very costly in terms of computational resources.