Artificial Intelligence (AI) is a science that deals with the problem of having machines perform intelligent, complex, actions with the aim of helping the human being. It is then possible to assert that Artificial Intelligence permits to bring into machines, typical characteristics and abilities that were once limited to human intervention. In the field of AI there are several tasks that ideally could be delegated to machines, such as environment aware perception, visual perception and complex decisions in the various field. The recent research trends in this field have produced remarkable upgrades mainly on complex engineering systems such as multi-agent systems, networked systems, manufacturing, vehicular and transportation systems, health care; in fact, a portion of the mentioned engineering system is discussed in this PhD thesis, as most of them are typical field of application for traditional control systems. The main purpose if this work is to present my recent research activities in the field of complex systems, bringing artificial intelligent methodologies in different environments such as in telecommunication networks, transportation systems and health care for Personalized Medicine. The designed and developed approaches in the field of telecommunication net- works is presented in Chapter 2, where a multi-agent reinforcement learning algorithm was designed to implement a model-free control approach in order to regulate and improve the level of satisfaction of the users, while the research activities in the field of transportation systems are presented at the end of Chapter 2 and in Chapter 3, where two approaches regarding a Reinforcement Learning algorithm and a Deep Learning algorithm were designed and developed to cope with tailored travels and automatic identification of transportation moralities. Finally, the research activities performed in the field of Personalized Medicine have been presented in Chapter 4 where a Deep Learning and Model Predictive control based approach are presented to address the problem of controlling biological factors in diabetic patients.

Le metodologie di Intelligenza Artificiale (AI) si occupano della possibilità di rendere le macchine in grado di compiere azioni intelligenti con lo scopo di aiutare l’essere umano; quindi è possibile affermare che l’Intelligenza Artificiale consente di portare all’interno delle macchine, caratteristiche tipiche considerate come caratteristiche umane. Nello spazio dell’Intelligenza Artificiale ci sono molti compiti che potrebbero essere richiesti alla macchina come la percezione dell’ambiente, la percezione visiva, decisioni complesse. La recente evoluzione in questo campo ha prodotto notevoli scoperte, princi- palmente in sistemi ingegneristici come sistemi multi-agente, sistemi in rete, impianti, sistemi veicolari, sistemi sanitari; infatti una parte dei suddetti sistemi di ingegneria è presente in questa tesi di dottorato. Lo scopo principale di questo lavoro è presentare le mie recenti attività di ricerca nel campo di sistemi complessi che portano le metodologie di intelligenza artifi- ciale ad essere applicati in diversi ambienti, come nelle reti di telecomunicazione, nei sistemi di trasporto e nei sistemi sanitari per la Medicina Personalizzata. Gli approcci progettati e sviluppati nel campo delle reti di telecomunicazione sono presentati nel Capitolo 2, dove un algoritmo di Multi Agent Reinforcement Learning è stato progettato per implementare un approccio model-free al fine di controllare e aumentare il livello di soddisfazione degli utenti; le attività di ricerca nel campo dei sistemi di trasporto sono presentate alla fine del capitolo 2 e nel capitolo 3, in cui i due approcci riguardanti un algoritmo di Reinforcement Learning e un algoritmo di Deep Learning sono stati progettati e sviluppati per far fronte a soluzioni di viaggio personalizzate e all’identificazione automatica dei mezzi trasporto; le ricerche svolte nel campo della Medicina Personalizzata sono state presentate nel Capitolo 4 dove è stato presentato un approccio basato sul controllo Deep Learning e Model Predictive Control per affrontare il problema del controllo dei fattori biologici nei pazienti diabetici.

Artificial Intelligence based multi-agent control system / Lisi, Federico. - (2019 Feb 26).

Artificial Intelligence based multi-agent control system

LISI, FEDERICO
26/02/2019

Abstract

Artificial Intelligence (AI) is a science that deals with the problem of having machines perform intelligent, complex, actions with the aim of helping the human being. It is then possible to assert that Artificial Intelligence permits to bring into machines, typical characteristics and abilities that were once limited to human intervention. In the field of AI there are several tasks that ideally could be delegated to machines, such as environment aware perception, visual perception and complex decisions in the various field. The recent research trends in this field have produced remarkable upgrades mainly on complex engineering systems such as multi-agent systems, networked systems, manufacturing, vehicular and transportation systems, health care; in fact, a portion of the mentioned engineering system is discussed in this PhD thesis, as most of them are typical field of application for traditional control systems. The main purpose if this work is to present my recent research activities in the field of complex systems, bringing artificial intelligent methodologies in different environments such as in telecommunication networks, transportation systems and health care for Personalized Medicine. The designed and developed approaches in the field of telecommunication net- works is presented in Chapter 2, where a multi-agent reinforcement learning algorithm was designed to implement a model-free control approach in order to regulate and improve the level of satisfaction of the users, while the research activities in the field of transportation systems are presented at the end of Chapter 2 and in Chapter 3, where two approaches regarding a Reinforcement Learning algorithm and a Deep Learning algorithm were designed and developed to cope with tailored travels and automatic identification of transportation moralities. Finally, the research activities performed in the field of Personalized Medicine have been presented in Chapter 4 where a Deep Learning and Model Predictive control based approach are presented to address the problem of controlling biological factors in diabetic patients.
26-feb-2019
File allegati a questo prodotto
File Dimensione Formato  
Tesi_dottorato_Lisi.pdf

accesso aperto

Tipologia: Tesi di dottorato
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 7.63 MB
Formato Adobe PDF
7.63 MB Adobe PDF

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11573/1259554
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact