A decision support system for vehicle routing based on model inversion and data analysis

Overview & Synopsis

In this paper, we present a three-phase solving mechanism for the vehicle routing problem. The solving mechanism is part of a decision support system architecture deduced from an interdisciplinary study. We highlight that human factors and dynamic aspects are generally ignored in the classical approaches to solve the problem. In our approach, a link is done between methods of operations research and an ecological interface design coming from cognitive ergonomics. We focus our study in how to manage the constraint relaxation if the problem is not satisfiable. We propose and evaluate model inversion techniques and data classification based methods in order to determine the most suitable constraints to relax in priority.