**Título:** Um Algoritmo Genético Para O Problema Dos K-Medoids

**Autores:** Brito, Wagner Marques de; Semaan, Gustavo Silva; Brito, José André de Moura

**Resumo:** This paper aims to present a new algorithm for a classical clustering problem, known as the k-medoids problem. In this problem, given a set of n objects with p attributes, you must allocate them to k distinct groups, so that each group is as homogeneous as possible, given an objective function that measures the distance (Euclidean) of each group object to a particular object (also in the group) called medoid. The high complexity associated with this problem led the study and implementation of an algorithm based on genetic algorithms metaheuristic. This metaheuristic is applied in many optimization problems, including various clustering problems. In order to assess the efficiency of this algorithm, several computational experiments were performed considering an artificial data repository and the data set of Brazilian HDI (Human Development Index).

**Palavras-chave:** Clustering; K-Medoids; Genetic Algorithms

**Páginas:** 8

**Código DOI:** 10.21528/CBIC2011-10.6

**Artigo em pdf:** st_10.6.pdf

**Arquivo BibTex:** st_10.6.bib