Optsicom. Optimization of Complex Systems

Description

This project deals with the development of optimization tools based on metaheuristic procedures. Its objective is the resolution of hard optimization problems and the development of a computer aided decision support system. Complex systems appear in different contexts, from the economy to the enginery and in most of the cases they combine optimization and simulation. In this project we will complete the software generated in the previous project for the optimization of different models with the objective of providing solutions to a wide range of theoretical and applied problems.

This project is built upon the strong and successful collaboration of two research groups. For more than 10 years, the group at the University of Valencia, coordinated by Prof. Dr. Rafael Martí, has actively collaborated with the group at the Rey Juan Carlos University, coordinated by Prof. Dr. Abraham Duarte. This collaboration started in 2006 and obtained its first output as a paper published in 2007 in the European Journal of Operational Research. Since then, both teams have published together almost 50 JCR-indexed papers, 25 book chapters, and their research has been presented in more than 50 international conferences. It has been funded by the Spanish Goverment (Ministerio de Ciencia, Innovación y Universidades: TIN2009-07516, TIN2012-35632, and TIN2015-65460), and has received remarkable evaluations.

We target two families of hard optimization problems: (1) those based on structured models, and (2) those in which the model is not completely known. In the former, in which a whole description of the problem is known, we face some of the most applied models in practice, proposing efficient solution methods based on meta-heuristic procedures. In the second case, in which the model is not completely defined, we will develop evolutionary methods that incorporate learning strategies for an efficient search of the solution space. In most of these problems, the evaluation of the objective function consists of a simulation procedure, resulting in an extremely time consuming operation. This project will have mainly two outputs: firstly a software program which will provide high quality solutions and secondly publications in the most prestigious journals of the area.