neo2016

Speakers - Carlos Segura

 

Carlos Segura
CIMAT, Mexico

Title: Management of Diversity in Evolutionary Algorithms and its application to the Frequency Assignment Problem

Abstract: Population-based metaheuristics have yielded very promising results in a large amount problems. However, there are several issues that affect their performance and that appear frequently in the design of population-based strategies. Two of the most common issues are premature convergence and the way of handling neutrality. One of the problems where these issues are affecting the performance of state-of-the-art strategies is the Frequency Assignment Problem. In this problem, metaheuristics have yielded very promising results. However, the results obtainable using currently published methods are far from ideal in complex, large-scale instances. In this talk, a review of the state-of-the-art in the application of metaheuristics to the frequency assignment problem will be briefly presented. Additionally, some new principles to design metaheuristics in a way that the previously mentioned drawbacks can be alleviated are discussed. A recently developed strategy that combines several of the most advanced operators designed for the frequency assignment problem with these design principles is presented. Specifically, the method applies a recent replacement-based diversity management strategy to face these two issues. The relation between the distance metric applied in this strategy and the way of handling neutrality is shown. The performance of the methods presented in this talk will be discussed by considering 44 publicly available instances, showing that new best-known solutions can be obtained in several cases. The application of similar principles to different problems will also be discussed.

Bio: Carlos Segura was born in Santa Cruz de Tenerife, Spain, on August 03, 1983. He received his M.S. degree in Computer Science from the Universidad de La Laguna in 2009, and his Ph.D. degree in Computer Science from the Universidad de La Laguna, in 2012. He was a teaching assistant in the University of La Laguna and a post-doc researcher in the Center for Research and Advanced Studies in Mexico D.F. He is currently an Associate Researcher in the Computer Science Department at the Centre for Research in Mathematics, in Guanajuato. Carlos Segura has authored and co-authored over 50 technical papers and book chapters, including 16 journal papers. His publications currently report over 300 citations in Google Scholar and his h-index is 13. Currently he serves on the editorial board of several international conferences. His main research interests are: design of evolutionary algorithms, diversity management and problem solving paradigms. Dr. Segura is a Member of the IEEE and a member of the ACM.