Simheuristic algorithms for the sustainable freight transport problem

• Autores: Lorena Reyes Rubiano
• Directores de la Tesis: Francisco Javier Faulín Fajardo (dir. tes.) , Ángel Alejandro Juan Pérez (codir. tes.)
• Lectura: En la Universidad Pública de Navarra ( España ) en 2019
• Idioma: inglés
• Tribunal Calificador de la Tesis: Fermín Mallor Giménez (presid.) , Carles Serrat i Piè (secret.) , Kris Braekens (voc.)
• Texto completo no disponible (Saber más ...)
• Resumen

  • The sustainable freight transport entails the design of the distribution plans with the least negative impacts. On one hand, this distribution problem relies on determining the routes to visit a set of customers, which can be geographically scattered. One the other hand, the operational constraints and the attributes involved in urban transport need to be considered for designing the distribution plan. Distribution plans encompass not only the classical routing constraints but also a set of economic, social and environmental criteria implicated with transport sector. These attributes link social and industrial needs taking into account the triple bottom of objectives sustainability. Those attributes may be difficult to address because they can be progressing in different directions. This thesis contributes to integrate these challenges by means of analysis of transport problems, and structured method developments for supporting the decision making process.

    To attain these challenges the following objectives have been proposed: • Identification of attributes and constrains for problems related to freight transport in smart cities, with especial focus on environmental, economic and social impacts.

    • Modeling of sustainability indicators in the vehicle routing problems with the purposes of producing greener transport in smart cities.

    • Design and implementation of hybrid algorithms combining metaheuristics with simulation to provide sustainable solutions.

    • Validation of the algorithms using realistic data and well-known solutions.

    The first objective is to provide a characterization in problems related to freight transport, considering a special focus on sustainability dimensions. Some measures to estimate the negative impacts caused by transport activities have been also included.

    In Chapter 1, the classical issues related to urban transport and the sustainability dimensions are presented.

    Afterwards, the Chapter 2 provides a general description of solving approaches for combinatorial optimization problems considering also an overview of the most common attributes and constraints related to the current sustainability initiatives. Then, the framework of biased randomized simheuristic algorithm is described together with the most classical methods to solve rich vehicle routing problems. The proposed algorithms are well described across the chapters of this thesis.

    For the second objective of this dissertation, a formal description for routing problems with single depot and multi depot configuration. In Chapter 3 a sustainable multi-depot problem is defined and solved by a mixed integer programming and a variable neighborhood search framework. From Chapter 4 to Chapter 6, vehicles routing problem with electric is described assuming a single depot and stochastic variables.

    The third objective is a global one which will be addressed over the course of the whole dissertation. Easy to implement and competitive simheuristic algorithms are proposed to cope with stochastic problems. Particular attention is paid on the inclusion of sustainable criteria and consideration of current operational constraints from freight transport.

    The fourth objective is to implement and test the algorithms using benchmarks for deterministic and stochastic problems. The results show the sustainability influence of the optimization criteria and the effect of stochastic data on the performance of the solution approaches and solutions quality. Finally, this dissertation ends with some conclusions and comments on further research lines.