A software framework based on a conceptual unified model for evolutionary multiobjective optimization: ParadisEO-MOEO

Abstract

International audienceThis paper presents a general-purpose software framework dedicated to the design and the implementation of evolutionary multiobjective optimization techniques: ParadisEO-MOEO. A concise overview of evolutionary algorithms for multiobjective optimization is given. A substantial number of methods has been proposed so far, and an attempt of conceptually unifying existing approaches is presented here. Based on a fine-grained decomposition and following the main issues of fitness assignment, diversity preservation and elitism, a conceptual model is proposed and is validated by regarding a number of state-of-the-art algorithms as simple variants of the same structure. This model is then incorporated into the ParadisEO-MOEO software framework. This framework has proven its validity and high flexibility by enabling the resolution of many academic, real-world and hard multiobjective optimization problems

MOCA-I: Discovering Rules and Guiding Decision Maker in the Context of Partial Classification in Large and Imbalanced Datasets

“…Moreover, DMLS is easier to parameter than a genetic algorithm as we only have to define a neighborhood operator. Therefore we used DMLS implemented by Liefooghe et al [16] in ParadisEO framework [17], with an unbounded archive, using the natural stopping criterion. DMLS algorithm is detailed in Algorithm 1.…”

“…We used Weka software version 3.6 for discretization of datasets and for running C4.5 tests. Our approach is implemented in C++, using metaheuristics from ParadisEO framework [17]. In our experimentations we set MOCA-I max ruleset size = 5, max rule size = 9 for each dataset.…”

Automatic Design of Evolutionary Algorithms for Multi-Objective Combinatorial Optimization

“…These works study MOEAs as monolithic units and provide insights on which particular MOEAs are state of the art for specific problems, giving a baseline for future developments. More recently, a component-wise view of MOEAs has drawn the attention of the MOEA community [5,13]. "Deconstructed" MOEAs actually differ by a few main algorithmic components, which can be individually analyzed to assess their actual contribution to the overall efficiency of the algorithm.…”

“…"Deconstructed" MOEAs actually differ by a few main algorithmic components, which can be individually analyzed to assess their actual contribution to the overall efficiency of the algorithm. This component-wise view has been recently strengthened by the development of flexible algorithmic frameworks [11,13], where novel MOEAs can be devised combining existing algorithmic components. However, the potential of such approach remains unclear as the number of possible combinations is extremely large to be fully explored.…”

“…In a first step, we present the framework we use for this work, which combines components from ParadisEO-MOEO [13], PISA [7], and PaGMO [6]. Particularly, some of the MOEAs from the literature cannot be easily represented just by the combination of a fitness and a diversity component as proposed in ParadisEO-MOEO.…”