96 research outputs found
A deep reinforcement learning framework for scalable slice orchestration in beyond 5G networks
No full textThe curse of (too much) choice: Handling combinatorial action spaces in slice orchestration problems using DQN with coordinated branches
No full text
Multi-agent DQN with sample-efficient updates for large inter-slice orchestration problems
No full text
The Daughter by Pavlos Matesis and by Pavel Kohout
Get PDFBachelor thesis is focused on modern Greek literature and its most important author of twentieth century Pavlos Matesis and presents his work. The central theme of the thesis is novel The Daughter. This paperwork puts this novel in to the context of the whole author's works and watches next transformations of the piece. One of the other forms of this novel is an adaptation written by Pavel Kohout. And the other is a dramatization by the author Pavlos Matesis himself
Un cadre d'apprentissage par renforcement profond pour l'orchestration de tranches évolutives dans les réseaux au-delà de la 5G
No full textThis Thesis introduces a flexible Reinforcement Learning queuing-based framework for dynamic slice orchestration in Beyond 5G networks, supporting multiple concurrent slices that span different technological domains and are governed by diverse end-to-end Service Level Agreements. Different (Deep) Reinforcement Learning methods (single or multi-agent) are investigated to address the state and action complexity hurdles arising in such combinatorial problems, which render the use of "vanilla" Reinforcement Learning algorithms impractical. The performance of the proposed schemes is validated through simulations under both synthetic Markovian traffic and real traffic scenarios.Cette thèse présente un cadre flexible basé sur l'apprentissage par renforcement des files d'attente pour l'orchestration dynamique des tranches dans les réseaux Beyond 5G, prenant en charge de multiples tranches concurrentes qui couvrent différents domaines technologiques et sont régies par divers accords de niveau de service de bout en bout. Différentes méthodes d'apprentissage par renforcement profond (mono ou multi-agents) sont étudiées pour résoudre les problèmes de complexité d'état et d'action liés à ces problèmes combinatoires, qui rendent l'utilisation d'algorithmes d'apprentissage par renforcement classique impraticable. La performance des schémas proposés est validée par des simulations dans des scénarios de trafic markovien synthétique et de trafic réel
Safety implications of higher levels of automated vehicles: a scoping review
No full textAutomated vehicles (AVs) promise to improve road safety, reduce traffic congestion and emissions, and enhance mobility. However, evidence regarding their safety benefits has not been systematically investigated and documented. In this study, we utilise a scoping review approach to investigate and synthesise the existing literature on higher levels of AVs’ safety implications. This aids future relevant studies by identifying the research gaps and reporting the methodological approaches used. The review focused not only on peer-reviewed articles but also on grey literature to provide a comprehensive overview of the current research state. In total, 5724 articles were identified, and 4167 records were screened after duplicates and dual publications removal, from which 27 were found eligible for review. Ultimately, 24 studies met all the inclusion criteria and were considered for the review. The reported evidence was focused on changes in road safety levels after the deployment of AVs in transport networks. The data was extracted and charted by one reviewer using tables to create a descriptive summary of the results and address the scoping review's questions and objectives. In general, the findings suggest that AVs hold the potential to improve the overall safety on roads, although the existing evidence is not mainly based on real data but assumptions regarding vehicles’ capabilities and behaviour. The limited number of studies and the fact that all of them were published or conducted after 2014 indicate that the research on AVs’ safety impacts is just emerging.Accepted Author ManuscriptTransport and Plannin
Un cadre d'apprentissage par renforcement profond pour l'orchestration de tranches évolutives dans les réseaux au-delà de la 5G
No full textCette thèse présente un cadre flexible basé sur l'apprentissage par renforcement des files d'attente pour l'orchestration dynamique des tranches dans les réseaux Beyond 5G, prenant en charge de multiples tranches concurrentes qui couvrent différents domaines technologiques et sont régies par divers accords de niveau de service de bout en bout. Différentes méthodes d'apprentissage par renforcement profond (mono ou multi-agents) sont étudiées pour résoudre les problèmes de complexité d'état et d'action liés à ces problèmes combinatoires, qui rendent l'utilisation d'algorithmes d'apprentissage par renforcement classique impraticable. La performance des schémas proposés est validée par des simulations dans des scénarios de trafic markovien synthétique et de trafic réel.This Thesis introduces a flexible Reinforcement Learning queuing-based framework for dynamic slice orchestration in Beyond 5G networks, supporting multiple concurrent slices that span different technological domains and are governed by diverse end-to-end Service Level Agreements. Different (Deep) Reinforcement Learning methods (single or multi-agent) are investigated to address the state and action complexity hurdles arising in such combinatorial problems, which render the use of "vanilla" Reinforcement Learning algorithms impractical. The performance of the proposed schemes is validated through simulations under both synthetic Markovian traffic and real traffic scenarios
- …
