Problemes de routing amb time windows aplicats a operacions de logística marítima
| Autor: | Farran Oliva, Jana |
|---|---|
| Přispěvatelé: | Universitat Politècnica de Catalunya. Departament d'Estadística i Investigació Operativa, Accenture, Vazquez Muiños, Henrique |
| Jazyk: | angličtina |
| Rok vydání: | 2023 |
| Předmět: |
time windows
Maritime transport Transport marítim heterogeneous fleet 90 Operations research mathematical programming::90B Operations research and management science [Classificació AMS] Matemàtiques i estadística [Àrees temàtiques de la UPC] pickup and delivery Algorismes adaptive large neighborhood search simulated annealing metaheuristic multiple materials Algorithms |
| Popis: | This project tackles the study of a rich variant of the Pickup and Delivery Problem with Time Windows (PDPTW) which aims to serve some customers requests to transport different amount of materials from one pickup location to its delivery location. The problem considers time windows for the pickups and deliveries and a heterogeneous fleet with multi-capacities for the different types of products. Moreover, we also consider a variant of the problem in a context of maritime transport logistics taking into account additional constraints related to maritime operations and defining an objective function with accurate operational costs. We present the design and implementation of a metaheuristic algorithm based on the Adaptive Large Neighborhood Search (ALNS) framework. The proposed method consists in iteratively destroying a part of the solution and reconstructing it by means of several specialized removal and insertion heuristics which are selected with an adaptive layer based on its historical performance. We test the algorithm against public benchmark data and some modified data which allows us to validate the metaheuristic efficiency by comparing the results obtained with the best known solutions for the benchmark data. Additionally, in the absence of test cases in the literature for the maritime variant of the PDPTW, we randomly generate synthetic problem instances and analyze the quality of the results obtained and the convergence of the algorithm. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|