Banca de QUALIFICAÇÃO: DANIEL NOBRE PINHEIRO

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
DISCENTE : DANIEL NOBRE PINHEIRO
DATA : 22/06/2016
HORA: 14:00
LOCAL: Sala 02 - DCA
TÍTULO:

The heterogeneous penalty-free p-median problem with fuzzy logic

PALAVRAS-CHAVES:

Heterogeneous p-Median Problem, Fuzzy Logic, Consumer Segmentation.

PÁGINAS: 50
GRANDE ÁREA: Engenharias
ÁREA: Engenharia Elétrica
RESUMO:

This work presents a model for the Penalty-free Heterogeneous p-Median Problem with fuzzy logic (PFHPM-fuzzy) and proposes a heuristic method with the Variable Neighborhood Search (VNS) metaheuristic to solve it. The PFHPM is a clustering problem based on the classic p-Median (PM) problem, which seeks to partition a set of objects in p groups and to define one object from each group to represent it. The PM problem, however, doesn't consider the heterogeneity in data, i.e., it assumes that the characteristics of the objects cannot change from the point of view of different observers. The Heterogeneous p-Median (HPM) problem helps to understand how individuals perceive a set of objects. It seeks to sort I individuals in G groups such that individuals in the same group must have similar perceptions about the objects' characteristics. Furthermore, it seeks to partition the objects in a different way according to the point of view of each group of individuals, which explicits the heterogeneity of data. This work proposes a fuzzy approach of the PFHPM problem, where one object can be split into different categories. For example, if one object poses characteristics present in two distinct categories, then instead of choosing only one category for the object, it could be 50% in the first category and 50% in the second category. This separation enables a more detailed study about how individuals perceive the objects. To solve this problem we use exact methods to find lower bounds and the VNS metaheuristic to find upper bounds.

MEMBROS DA BANCA:
Presidente - 1746084 - DANIEL ALOISE
Externo ao Programa - 1746062 - CAROLINE THENNECY DE MEDEIROS ROCHA
Externo ao Programa - 2993322 - ÉVERTON SANTI