Heuristic Search Strategies for Noisy Optimization

Many real-world optimization problems are subject to noise, and making correct comparisons between candidate solutions is not straightforward. In the literature, various heuristics have been proposed to deal with this problem. Most studies compare evolutionary strategies with algorithms which propose candidate solutions deterministically. This paper compares the efficiency of different randomized heuristic search strategies, and also extends randomized algorithms non based on populations with a statistical analysis technique in order to deal with the presence of noise. Results show that this extension can outperform population-based algorithms, especially with higher levels of noise.