Qualificação de doutorado do aluno Vinícius Gandra, dia 05/07/2021, as 13 horas.
Qualificação de doutorado do aluno Vinícius Gandra, dia 05/07/2021, as 13 horas.
Título: Vehicle Routing Problems in Logistic Companies
Resumo:
The vehicle routing problem (VRP) was first introduced nearly 60 years ago and delineates a class of problems which remains an active research area of operations research. In simple terms, the VRP aims to minimize the cost of delivery routes from a central depot to a set of geographically scattered customers using a fleet of homogeneous vehicles subject to capacity constraints. The VRP is an NP-hard problem and thus poses a challenge to both exact and heuristic approaches.
Despite its complexity, the VRP in its original form is unable to capture the level of detail which characterizes real-world applications where various operational rules are encountered. In order to bridge the gap between academic theory and practical applications the original VRP is often expanded with real-world constraints, leading to a range of computationally challenging VRP variants.
The research presented in this thesis is motivated by distribution challenges arising in logistic companies which have not yet been completely captured in existing problems. This thesis introduces VRP generalizations which combine multiple constraints occurring at different decision levels to tackle these real-world problems. These levels of decisions include, besides the routing, the location decisions concerning where to transfer goods and loading decisions which aim to produce feasible two-dimensional load plans. Such problem generalizations require more decisions to be taken into account, significantly expanding the search space and posing a great challenge for companies which currently rely on off-the-shelf solvers or human operators to produce solutions. To obtain high-quality solutions in short computational runtimes, heuristics are developed which comprise of tailored methods capable of efficiently handling problem-specific constraints and which deploy state-of-the-art algorithms to solve the VRP component of the problems. This thesis also investigates the impact of the newly introduced constraints on solution quality. Computational experiments are performed to evaluate the quality of the proposed methods and solutions obtained using instances derived from real-world data which have been made publicly available to stimulate further research.
Banca: Prof. Dr. Anand Subramanian (UFPB), Prof. Dr. Marcone Freitas Souza (UFOP), Prof. Dr. Puca Vaz Penna (UFOP)
Data: 05/07/2021 (Segunda-feira)
Horário: 13hrs
Link: https://meet.google.com/fsy-vuco-wro
