Diseño metodológico para integración de investigación de operaciones e inteligencia artificial: apoyo para modelos de localización-asignación en logística alimentaria

Abstract

This monograph proposes a methodological design for the integration of operations research (linear, integer linear and mixed linear programming) and artificial intelligence (clustering, regression, and classification algorithms) to support allocation-location models in food logistics. Using a case study to evaluate the existence of a synergy between the two disciplines and their ability to reduce time and computational resources with a good response quality. A historical review of both branches is made looking for similarities, defining the key aspects in which they coincide in the academic documentation found. Defining also, the inputs to be used and evaluated. The existence of four structures in the synergy is proposed: value generation, data training, parameter selection and problem division. Subsequently, they are evaluated in the case of food logistics as a test space, where the location-assignment of distribution centers for the last step of the supply chain in Colombia is experimented with the CFLP (capacitated facility location problem) model, regarding the municipalities connected by land routes. Finally, an analysis is made from the application of the workflow establishing the weaknesses and strengths through an approach of costs and projection within the national logistics network. Promising results were found in terms of the use of computational resources, time and costs by making use of the union between disciplines.