Output details
11 - Computer Science and Informatics
University of Birmingham
Analysis of Computational Time of Simple Estimation of Distribution Algorithms
<22>This was the first paper on the rigorous analysis of computational time complexity of estimation of distribution algorithms (EDAs), which is very different from conventional evolutionary algorithms. A new approach to proving complexity bounds for randomised search heuristics, especially for EDAs, was proposed. The analytical approach and results have enabled us to understand theoretically which algorithmic feature(s) makes a problem hard/easy for an EDA, and given insight into the question of when an EDA is expected to perform better than an EA. The research led to invited keynote speeches on this topic at LION'10 in Italy and WCCI'12 in Brisbane.