Histogram-Based Estimation of Distribution Algorithm: A Competent Method for Continuous Optimization

Designing efficient estimation of distribution algorithms for optimizingcomplex continuous problems is still a challenging task. This paperutilizes histogram probabilistic model to describe the distribution ofpopulation and to generate promising solutions. The advantage ofhistogram model, its intrinsic multimodality, makes it proper to describe the solution distribution ofcomplex and multimodal continuous problems. To make histogram modelmore efficiently explore and exploit the search space, severalstrategies are brought into the algorithms: the surrounding effectreduces the population size in estimating the model with a certainnumber of the bins and the shrinking strategy guarantees the accuracyof optimal solutions. Furthermore, this paper shows thathistogram-based EDA (Estimation of distribution algorithm) can givecomparable or even much better performance than those predominant EDAsbased on Gaussian models.