This research work focuses on the study of dynamic modeling and the role that uncertainty plays within the various complex decision-making problems. Complexity theory deals with analyzing the computational complexity of an algorithm. The main goal was to elaborate the most efficient (to our best knowledge) algorithms to solve the proposed problems. Therefore, the use of innovative Machine Learning techniques allows you to automate the learning process and autonomously provide scenarios and solutions to support decision-making processes. The value and reliability of data-based decisions depend on the quality of the data, the correct analysis of the same and the right interpretations. However, this cannot happen in the absence of a correct use of Big Data and Machine Learning techniques. The present Ph.D. thesis developed jointly with an international research group focuses on the development of mathematical optimization models and new Machine Learning and Deep Learning algorithms for making predictions and analyzing social dynamics, in particular pandemic and financial issues. The obtained results have shown the efficiency and robustness of the new approaches to address the conditions of uncertainty. In this work new findings were gained in the frame of (fuzzy) fractional calculus, stochastic programming, mathematical programming, machine and deep learning. A fruitful and rigorous combination among mathematics and artificial intelligence
Questo lavoro di ricerca si concentra sullo studio della modellazione dinamica e sul ruolo che l'incertezza gioca all'interno dei vari problemi decisionali complessi. La teoria della complessità si occupa dell'analisi della complessità computazionale di un algoritmo. L'obiettivo principale è stato quello elaborare gli algoritmi più efficienti (a nostra conoscenza) per risolvere i problemi proposti. Pertanto, l'utilizzo di tecniche innovative di Machine Learning consente di automatizzare il processo di apprendimento e fornire autonomamente scenari e soluzioni a supporto dei processi decisionali. Il valore e l'affidabilità delle decisioni basate sui dati dipendono dalla qualità dei dati, dalla corretta analisi degli stessi e dalle giuste interpretazioni. Tuttavia, ciò non può avvenire in assenza di un corretto utilizzo delle tecniche di Big Data e Machine Learning. La presente tesi di dottorato di ricerca sviluppata in collaborazione con un gruppo di ricerca internazionale si concentra sullo sviluppo di modelli matematici di ottimizzazione e nuovi algoritmi di Machine Learning e Deep Learning per fare previsioni e analizzare le dinamiche sociali, in particolare, le problematiche pandemiche e finanziarie. I risultati ottenuti hanno mostrato l'efficienza e la robustezza dei nuovi approcci per affrontare le condizioni di incertezza. In questo lavoro sono state ottenute nuove scoperte nell'ambito del calcolo frazionario (fuzzy), della programmazione stocastica, della programmazione matematica, del Machine e Deep Learing. Un connubio fruttuoso e rigoroso tra matematica e Intelligenza Artificiale
Advances on dynamics modeling: new iusses on mathematical programming, artificial intelligence and Covid-19 transmission forecasting / Ciano, Tiziana. - (2022 Oct 18).
Advances on dynamics modeling: new iusses on mathematical programming, artificial intelligence and Covid-19 transmission forecasting
Ciano, Tiziana
2022-10-18
Abstract
This research work focuses on the study of dynamic modeling and the role that uncertainty plays within the various complex decision-making problems. Complexity theory deals with analyzing the computational complexity of an algorithm. The main goal was to elaborate the most efficient (to our best knowledge) algorithms to solve the proposed problems. Therefore, the use of innovative Machine Learning techniques allows you to automate the learning process and autonomously provide scenarios and solutions to support decision-making processes. The value and reliability of data-based decisions depend on the quality of the data, the correct analysis of the same and the right interpretations. However, this cannot happen in the absence of a correct use of Big Data and Machine Learning techniques. The present Ph.D. thesis developed jointly with an international research group focuses on the development of mathematical optimization models and new Machine Learning and Deep Learning algorithms for making predictions and analyzing social dynamics, in particular pandemic and financial issues. The obtained results have shown the efficiency and robustness of the new approaches to address the conditions of uncertainty. In this work new findings were gained in the frame of (fuzzy) fractional calculus, stochastic programming, mathematical programming, machine and deep learning. A fruitful and rigorous combination among mathematics and artificial intelligenceFile | Dimensione | Formato | |
---|---|---|---|
Ciano Tiziana.pdf
accesso aperto
Tipologia:
Tesi di dottorato
Licenza:
DRM non definito
Dimensione
6.42 MB
Formato
Adobe PDF
|
6.42 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.