A decision-making tool for the determination of the distribution center location in a humanitarian logistics network

Abstract

The distribution of humanitarian aid is a vital issue for humanity's future. In recent years, the management of humanitarian crises has become more crucial than it was a decade ago. Due to the volatility and urgency that characterize such situations, one of the most important challenges globally is the optimization of decisions regarding the timely distribution of aid during humanitarian operations. Our main goal is to develop an innovative decision-making tool, essential for non-profit organizations and governments that aims at the prompt selection of the location of the distribution center of humanitarian aid, in cases of natural or human-made disasters. The proposed tool is based on network science principles and can be used for selecting a suitable node for the installation of a distribution center during the beginning of a humanitarian crisis, considering that networks have a volatile nature and require quick decisions. For the configuration of the proposed tool we use a combination of a classical heuristic algorithm and predictive models based on a binary classification problem with the support of a supervised deep neural network. It is developed using the R programming language with the contribution of the “Shiny” package (web application framework for R) along with other packages for network analysis, data manipulation and visualization.