28. A Scalable Evolutionary Algorithm with Intensification and Diversification Protocols Designed for Statistical Models
Authors: Wendy K. Cho (University of Illinois)Yan Y. Liu (University of Illinois)
Abstract: Important insights into many problems that are traditionally analyzed via statistical models can be obtained by re-formulating within a large-scale optimization framework. The theoretical underpinnings of statistical models may shift the goal of the solution space traversal from the traditional search for an optimal solution to a traversal with the purpose of yielding a set of high quality, independent solutions. We examine statistical frameworks with astronomical solution spaces where the independence requirement constitutes a significant additional challenge for standard optimization methodologies. We design a hybrid metaheuristic with intensification and diversification protocols in the base search algorithm. Via our grant on the Blue Waters supercomputer, we extend our algorithm to the high-performance-computing realm. We experimentally demonstrate the effectiveness of our algorithm to utilize multiple processors to collaboratively hill climb, broadcast messages to one another about the landscape characteristics, diversify across the landscape, and request aid in climbing particularly difficult peaks.
Two-page extended abstract: pdf