Multiobjective estimation distribution algorithm for continuous, mixed and overlapping problems

• Grant number: 12/17088-5
• Support Opportunities: Scholarships abroad - Research Internship - Doctorate
• Start date: October 01, 2012
• End date: December 31, 2012
• Field of knowledge: Physical Sciences and Mathematics - Computer Science
• Principal Investigator: Alexandre Cláudio Botazzo Delbem
• Grantee: Antonio Helson Mineiro Soares
• Supervisor: António Gaspar Lopes da Cunha
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
• Institution abroad: Universidade do Minho (UMinho), Portugal
• Associated to the scholarship: 10/01705-0 - Estimation of Distribution Algorithms based on Phylogenetic Trees, BP.DR

Abstract

The \PhiGA, developed by the research group of LCR-ICMC/USP was doubly inspired by principles of the Theory of Evolution: (i) by using the concepts of the generation of diversity and solution as well as the evolutionary algorithms; (ii) to explain a set of individuals using relationships between characteristics of these individuals through a phylogenetic tree. Phylogenies can be understood as probabilistic models similar to the estimation distribution algorithms that instead of correlated variables determines relationships (genealogical similarities) between species. The mo\PhiGA, multiobjective extension of \PhiGA, overcame the commonly used algorithms in the literature for deceptivos decomposable problems. Based on these algorithms was developed Net\PhiGA and moNet\PhiGA which has even better results for that class of problems. This research project in exterior aims to increase the set of problems that the new algorithm can solve. (AU)