DistrictNet: Decision-aware learning for geographical districting

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Ahmed, Cheikh, Forel, Alexandre, Parmentier, Axel, Vidal, Thibaut
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866929624795054080
author Ahmed, Cheikh
Forel, Alexandre
Parmentier, Axel
Vidal, Thibaut
author_facet Ahmed, Cheikh
Forel, Alexandre
Parmentier, Axel
Vidal, Thibaut
contents Districting is a complex combinatorial problem that consists in partitioning a geographical area into small districts. In logistics, it is a major strategic decision determining operating costs for several years. Solving districting problems using traditional methods is intractable even for small geographical areas and existing heuristics often provide sub-optimal results. We present a structured learning approach to find high-quality solutions to real-world districting problems in a few minutes. It is based on integrating a combinatorial optimization layer, the capacitated minimum spanning tree problem, into a graph neural network architecture. To train this pipeline in a decision-aware fashion, we show how to construct target solutions embedded in a suitable space and learn from target solutions. Experiments show that our approach outperforms existing methods as it can significantly reduce costs on real-world cities.
format Preprint
id arxiv_https___arxiv_org_abs_2412_08287
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle DistrictNet: Decision-aware learning for geographical districting
Ahmed, Cheikh
Forel, Alexandre
Parmentier, Axel
Vidal, Thibaut
Machine Learning
Optimization and Control
Districting is a complex combinatorial problem that consists in partitioning a geographical area into small districts. In logistics, it is a major strategic decision determining operating costs for several years. Solving districting problems using traditional methods is intractable even for small geographical areas and existing heuristics often provide sub-optimal results. We present a structured learning approach to find high-quality solutions to real-world districting problems in a few minutes. It is based on integrating a combinatorial optimization layer, the capacitated minimum spanning tree problem, into a graph neural network architecture. To train this pipeline in a decision-aware fashion, we show how to construct target solutions embedded in a suitable space and learn from target solutions. Experiments show that our approach outperforms existing methods as it can significantly reduce costs on real-world cities.
title DistrictNet: Decision-aware learning for geographical districting
topic Machine Learning
Optimization and Control
url https://arxiv.org/abs/2412.08287