Scalable optimization via probabilistic modeling

Más información sobre el libro

I’m not usually a fan of edited volumes, as they often present a disjointed collection of articles under misleading titles. However, this volume is a commendable exception, successfully focusing on a specific and relevant topic: estimation of distribution algorithms (EDAs). These algorithms combine evolutionary computation’s population orientation and selectionism, discarding genetics to create a powerful and elegant hybrid. Unlike many edited collections, the articles here are logically sequenced, guiding the reader from design to efficiency enhancement and concluding with practical applications. The focus on efficiency is particularly noteworthy, as the data-mining perspective inherent in EDAs introduces new methods for data-guided adaptation. This approach can significantly accelerate solutions by leveraging effective surrogates, hybrids, and parallel and temporal decompositions. Overall, this book stands out for its coherence and relevance, making it a valuable addition to any library interested in cutting-edge optimization techniques.