| Grant number: | 19/22067-6 |
| Support Opportunities: | Scholarships abroad - Research |
| Start date: | March 05, 2020 |
| End date: | October 04, 2020 |
| Field of knowledge: | Engineering - Production Engineering - Operational Research |
| Principal Investigator: | Mariá Cristina Vasconcelos Nascimento Rosset |
| Grantee: | Mariá Cristina Vasconcelos Nascimento Rosset |
| Host Investigator: | Jean-François Cordeau |
| Host Institution: | Instituto de Ciência e Tecnologia (ICT). Universidade Federal de São Paulo (UNIFESP). Campus São José dos Campos. São José dos Campos , SP, Brazil |
| Institution abroad: | École des Hautes Études Commerciales (HEC Montréal), Canada |
| Associated research grant: | 13/07375-0 - CeMEAI - Center for Mathematical Sciences Applied to Industry, AP.CEPID |
Abstract Deep learning (DL) has been successfully applied to a variety of signal and information processing tasks with outstanding results. It consists of computational models composed by multiple processing layers whose goal is to learn data representation considering multiple levels of abstraction. Several recent attempts to address combinatorial optimization problems (COPs) by deep networks can be observed in the literature. In spite of the efforts, to develop models competitive with state-of-the-art heuristic methods remains a challenge. One of the major problems of DL for COPs, mainly in the end-to-end learning solutions, not yet overcome is the scalability. Therefore, this project proposes incorporating deep learning in a local search-based metaheuristic. For such, a linear-time deep learning strategy based on neighborhoods in graphs will be investigated. The goal is to achieve scalability and obtain competitive results in tackling routing problems. (AU) | |
| News published in Agência FAPESP Newsletter about the scholarship: | |
| More itemsLess items | |
| TITULO | |
| Articles published in other media outlets ( ): | |
| More itemsLess items | |
| VEICULO: TITULO (DATA) | |
| VEICULO: TITULO (DATA) | |