Egilea: Imanol Echeverria Franco

Izenburua: Deep learning for real time seheduling

Zuzendaria: Roberto Santana Hermida

Eguna: 2026ko ekainaren 15ean
Ordua: 10:30h
Lekua: Ada Lovelace aretoan (Informatikako fakultatea)

Abstract:

"This doctoral thesis explores the application of deep learning techniques to real-time combinatorial optimization problems, with a particular focus on industrial scheduling. The research investigates how modern machine learning approaches, such as graph neural networks, reinforcement learning, and hybrid optimization methods, can improve decision-making in complex and dynamic environments where traditional optimization techniques often struggle due to computational limitations. The thesis proposes new methodologies for representing scheduling problems, generating efficient solutions, and combining learning-based approaches with classical optimization techniques. Through extensive experimentation on benchmark problems, the work demonstrates the potential of deep learning to produce scalable, adaptive, and high-quality solutions for real-time optimization, contributing to the advancement of intelligent decision-making systems in industrial applications."