Athens University Economics and Business (AUEB)
Optimization methods for combined problems in transportation and logistics management
Abstract
dc:descriptionThe globalization of the markets and the growing interdependence of the world's economies emphasize the role of supply chain management. The great majority of products follow a long journey before reaching the final customer. For instance, the production of the various pieces of a product and their assembly may take place in geographically different regions, whereas the wholesalers, the retailers and the customers of the final product may be located all over the planet. The efficient design and management of an organization's supply chain network constitutes a significant competitive advantage and determinant factor for the sustainability and prosperity of the organization. The challenges arising in the context of supply chain management, and especially in logistics management are formulated as complex combinatorial optimization problems that model various operations, such as production, inventory control and routing. Due to their high complexity, the dominant decision-making practice is to handle each operation separately ignoring their obvious interdependence to avoid the associated extremely complex management and optimization problems. However, relatively recent technological advances, as well as, academic research have opened the road for integration of supply chain operations. The pursue of bridging academic knowledge and practice and the advanced computational capabilities (e.g., cloud computing, parallel computing, computer clusters) guide the research towards integrated transportation and logistics optimization.The trend for integrated optimization is driven by the economic advantages that are achieved via integrated policies when traditionally separate procedures are jointly taken into account (e.g., location, production, inventory control, distribution, packing, routing, etc.). In all above cases, compared to the "traditional" sequential decision-making policies, the optimization of the operations in an integrated manner provides more benefits and reduces the costs. The emerging integrated optimization problems are modeled as combinatorial optimization problems and are mainly NP-hard problems. Their great importance both from the theoretical and commercial perspectives have attracted the attention of both practitioners and researchers from various fields of science, namely Operations Research, Mathematics, Computer Science, and Management Science. For the integrated optimization problems, the exhaustive search of the solution space is intractable even for small sized instances. Therefore, a wide range of exact, metaheuristic and matheuristic methodologies is employed to tackle these problems.In the present PhD thesis, a collection of complex operations research problems is examined and solved. The main characteristic of the studied models is that they integrate the traditional vehicle routing decisions with other procedures such as production decisions, inventory control and distribution decisions. In other words, these models integrate the day-to-day operational decisions with tactical decisions creating ample space for high-quality performance and cost savings. They arise in practical applications of transportation and logistics, and therefore, are of great importance for both practitioners and researchers. This thesis presents a spectrum of high performing methodological approaches (exact, metaheuristic, matheuristic) and also extends existing models to create more applied and realistic variants that better express common business requirements. Initially, the focus is on the Inventory Routing Problem (IRP) with Maximum Level inventory policy. The IRP is a broad class of hard-to-solve problems with numerous practical applications in the field of freight transportation and logistics (liquefied natural gas, fuel, petroleum products, agricultural products, fish, etc.). The generic IRP description involves a supplier responsible for determining the timing and the quantity of replenishment services offered to a set of customers over a multi-period time horizon. In addition, vehicle routes have to be defined jointly with the inventory related decisions. Due to the complexity of the problem, the IRP literature is relatively scarce in exact methods. To this end, a novel two-commodity flow formulation is introduced together with a new set of valid inequalities. Compared to known formulations, the nature of the formulation implies efficient scaling with the size of the vehicle fleet. On this basis, a branch-and-cut algorithm that employs methods for separating various well-known families of cuts is proposed. Extensive computational experiments are reported on well-established benchmark data sets. The proposed solution approach outmatches results of current state-of-the-art branch-and-cut, branch-and-price, metaheuristic and mathematical programming based heuristic approaches, especially for hard-to-solve instances. Secondly, the well-known Production Routing Problem (PRP) is studied. The PRP is considered an extension and a generalization of the IRP, since it introduces the production aspect to the decision-making process. Similarly, the PRP is a hard-to-solve combinatorial optimization problem with numerous practical applications in the field of supply chain management. A manufacturer is responsible for determining production decisions, as well as the timing and the quantity of replenishment services offered to a set of geographically dispersed customers over a multi-period time horizon. The problem calls for jointly optimizing the production, inventory, distribution and routing decisions. Building upon the novel IRP two-commodity flow formulation, a similar two-commodity flow formulation is proposed for the problem. However, the extreme complexity of the PRP dictates alternative methods, since exact ones cannot tackle even medium-sized instances. Toward this direction, a two-phase infeasible space matheuristic algorithm is developed. The first phase deals with a relaxation of the problem to construct production-distribution plans. In the second phase, these are completed with routing information and optimized via a local search framework which oscillates between the feasible and the infeasible solution space. The framework is equipped with mixed integer programming components for restoring infeasibility and for diversifying the conducted search. Computational experiments demonstrate that the infeasibility space exploration significantly contributes to the quality of the final solution. The results outperform the best known solution in the PRP literature for over 50% of the instances of two widely used benchmark data sets.Finally, a realistic variant of the PRP is introduced. The Periodic Production Routing Problem (PPRP) extends the PRP by incorporating periodicity requirements for the optimal decisions. The basic PRP definition focuses on a specific time-horizon creating a situation in which all customers and the production facility have zero inventory at the end of the studied time horizon. Therefore, during the first period after the horizon end, stock outs are inevitable for many customers. On this basis, the basic PRP model is modified to generate repeatable periodic production and delivery schedules that can be applied in cycles. A two-commodity flow formulation is proposed along with valid inequalities. Extensive comparisons between the basic PRP and the proposed periodic variant on well-known instances are provided. The new variant is significantly harder to solve, especially when the vehicle fleet is limited. From a managerial perspective, the generation of periodic production-routing schedules significantly increases all cost components: customers are forced to preserve higher inventory levels, whereas the number of vehicle routes required to implement a periodic schedule is approximately doubled. Concluding, this thesis focuses on inventory and production routing problems arising in the field of transportation logistics and supply chain management. Namely, the Inventory Routing Problem and the Production Routing problem are studied. A spectrum of powerful and efficient optimization methods is developed and proposed ranging from a Branch-and-Cut algorithm and Tabu Search to a hybrid matheuristic algorithm that incorporates mathematical programming components into a local search framework. Moreover, from a modelling aspect efficient two-commodity formulations are developed, whereas a new problem variant is introduced in pursue of reducing the gap between theoretical knowledge and practice.
Degree
thesis:*- Grantor dc:publisher
- Athens University Economics and Business (AUEB)
- Year dc:date
- 2021
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Manousakis, Eleftherios
- Μανουσάκης, Ελευθέριος
Subjects
dc:subject × 28- Διοίκηση εφοδιαστικής αλυσίδας
- Διοίκηση μεταφορών
- Συνδυαστική βελτιστοποίηση
- Μεταευρετικοί αλγόριθμοι
- Μαθ-ευρετικοί αλγόριθμοι
- Επιχειρησιακή έρευνα
- Υπολογιστική ευφυΐα
- Supply chain management
- Logistics management
- COMBINATORIAL OPTIMIZATION
- Metaheuristic algorithms
- Matheuristic algorithms
- Operations research
- Computational intelligence
- Μαθηματικά
- Φυσικές Επιστήμες
- Διακριτά μαθηματικά και Συνδυαστική
- Οικονομικά και Επιχειρήσεις
- Κοινωνικές Επιστήμες
- Διοικητική Επιστήμη και Επιχειρησιακή Έρευνα
- Επιστήμη λήψης αποφάσεων
- Mathematics
- Natural Sciences
- Discrete Mathematics and Combinatorics
- Economics and Business
- Social Sciences
- Management Science and Operations Research
- Decision Sciences
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/50314