Heurística de Busca de Vizinhança Variável para Otimização do Problema de Roteamento de Veículos

Resumo

ABSTRACT
This work investigates a computational method for the
optimization of the Enabled Vehicle Routing Problem. To propose
the method, the variable neighborhood search meta-heuristic
(Variable Neighborhood Search-VNS) was used. The proposed
heuristic contains the implements of 5 neighborhood structures, 4
of them built with deterministic methods, to explore the search
space in order to obtain local minima. In addition to these
strategies, a search with random factors is also presented, to try to
escape from local minima. For the computational tests, 4 groups
of instances were used, with different characteristics, seeking to
evaluate the heuristic in different scenarios. In the results it was
found that the heuristic was able to obtain good results, with an
error of up to 8% of the solution.