Abstract

PEREZ-SOSA, Teresa; CRUZ, Yarens; LA FE, Iván  and  QUIZA, Ramón. Class library for multi-objective optimization by using the cross-entropy method. Rev cuba cienc informat [online]. 2019, vol.13, n.1, pp. 90-104. ISSN 2227-1899.

Multi-objective optimization plays a key role in modern engineering practice, as it allows designing systems and processes where several criteria are simultaneously improved. In this field, it should be highlighted the use of the so-called a posteriori approach, which is based on obtaining the non-dominated solutions (also known as the Pareto set) for, after that, choosing from them the most convenient alternative for the actual conditions. It has emerged as the most effective choice, as it avoids the previous supply of any preference indication with respect to the objectives. This work presents the implementation of a class library for multi-objective optimization by using the cross-entropy method. It is based in a previous MATLAB toolbox developed and published by the authors. The library was implemented in C++, avoiding the use of any library but the standard ones, for guarantying the portability and integration capability. It will be distributed under the GNU Lesser General Public License (version 3). For testing not only the behavior of class library but also its insertion in other programs, an application was developed. All the tests carried out to the software have shown a good performance without identified mistakes. As a future development of this work, it has been foreseeing the implementation of other multi-objective optimization heuristic, such as genetic algorithms and ant colony optimization. It should be also integrated this library in applications for solving actual industrial problems, such as optimization of sustainability of manufacturing processes and design of micro-industries for food processing.

Keywords : Class library; cross-entropy; free software; multi-objective optimization; pareto front.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )