Multi Objective Ant Colony Optimisation to obtain efficient metro speed profiles

[EN] Obtaining efficient speed profiles for metro trains is a multi- objective optimisation problem where energy consumption and travel time must be balanced. Automatic Train Operation (ATO) systems may handle a great number of possible speed profiles; hence optimisation algorithms are required find...

Descripción completa

Detalles Bibliográficos
Autores: Martínez Fernández, Pablo|||0000-0002-8246-2510, Villalba Sanchis, Ignacio|||0000-0002-4091-8719, Insa Franco, Ricardo|||0000-0002-6655-4458, Font Torres, Juan B.
Tipo de recurso: artículo
Fecha de publicación:2023
País:España
Institución:Universitat Politècnica de València (UPV)
Repositorio:RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Idioma:inglés
OAI Identifier:oai:riunet.upv.es:10251/197954
Acceso en línea:https://riunet.upv.es/handle/10251/197954
Access Level:acceso abierto
Palabra clave:Energy efficiency
Optimization
Genetic algorithm
Ant colony
Metro trains
INGENIERIA E INFRAESTRUCTURA DE LOS TRANSPORTES
09.- Desarrollar infraestructuras resilientes, promover la industrialización inclusiva y sostenible, y fomentar la innovación
Descripción
Sumario:[EN] Obtaining efficient speed profiles for metro trains is a multi- objective optimisation problem where energy consumption and travel time must be balanced. Automatic Train Operation (ATO) systems may handle a great number of possible speed profiles; hence optimisation algorithms are required find efficient ones in a timely manner. This paper aims to assess the performance of a particular meta-heuristic optimisation algorithm, a variation of the traditional Ant Colony (ACO) modified to deal with multi-objective problems with continuous variables: MOACOr. This algorithm is used to obtain efficient speed profiles in up to 32 interstation sections in the metro network of Valencia (Spain), and the convergence and diversity of these solution sets is evaluated through metrics such as Inverse Generational Distance (GD) and Normalised Hypervolume (NH). The results are then compared to those obtained with a conventional genetic algorithm (NSGA-II), including a statistical analysis to identify significant differences. It has been found that MOACOr shows a better performance than NSGA-II in terms of convergence, regularity and diversity of the solution. These results indicate that MOACOr is a good alternative to the widely used genetic algorithm and could be a better tool for rail operation managers trying to improve energy efficiency.