Complex systems typically display emergent dynamic behaviors, which cannot be outlined by a mere description of their constituting parts. Such systems are ubiquitous in natural and artificial settings, ranging from biology and physics, to engineering and economics, and spark the interest of many scientists belonging to different research fields. To gain better insights on their inner working, complex systems are studied by means of computational methods, which allow to model and reproduce their behavior in silico. As a consequence, several mathematical formalisms were developed in the last decades to model complex systems, capture different features of their dynamic behavior, and leverage any available quantitative or qualitative data about their components. The aim of this thesis is to show how fuzzy logic can be exploited to overcome some of the limitations that still affect the modeling of complex systems, namely: predict emergent dynamic behaviors even when there is a lack of precise quantitative information; deal with the presence of heterogeneous systems components, spanning different levels of temporal, spatial of functional organization; bridge the gap between quantitative and qualitative modeling, in order to define hybrid models that can simultaneously exploit the peculiar advantages provided by both approaches. The novel modeling frameworks presented in this thesis based on fuzzy logic have been employed to analyze the behavior of two real world systems, in the context of cellular biology. The results show that the developed frameworks correctly reproduce the behavior of complex systems and assess their response to perturbations, even when quantitative information is missing or some system components are not fully characterized. These frameworks could find applications in several fields, including, but not limited to, biology, medicine and pharmacology, which often face such challenges.
I sistemi complessi generalmente mostrano comportamenti emergenti dinamici, che non possono essere spiegati da una mera descrizione delle parti che li costituiscono. Tali sistemi sono ubiquitari in contesti sia naturali che artificiali, spaziando dalla biologia alla fisica, dall’ingegneria all’economia, e suscitano l’interesse di molti scienziati appartenenti a diversi campi di ricerca. Per ottenere una più chiara comprensione del loro funzionamento, i sistemi complessi sono studiati tramite l’uso di metodi computazionali, che permettono di riprodurne il comportamento in silico. Di conseguenza, negli ultimi decenni numerosi formalismi matematici sono stati sviluppati per modellare i sistemi complessi, rappresentare diversi aspetti del loro comportamento dinamico e sfruttare i dati quantitative e qualitativi a disposizione riguardo le loro componenti. Lo scopo di questa tesi è mostrare come è possibile sfruttare la logica fuzzy per superare alcune delle limitazioni ancora presenti nella modellazione dei sistemi complessi, ovvero: predire comportamenti dinamici emergenti anche in assenza di informazioni quantitative precise; tener conto della presenza di componenti eterogenee, distribuite su diversi livelli di organizzazione temporale, spaziale e funzionale; colmare il divario esistente tra la modellazione quantitativa e quella qualitativa, in modo da definire modelli ibridi in grado di sfruttare simultaneamente i vantaggi peculiari di ciascun approccio. I nuovi framework di modellazione basati sulla logica fuzzy presentati in questa tesi sono stati utilizzati per analizzare il comportamento di due sistemi reali, nel contesto della biologia cellulare. I risultati dimostrano che i framework sviluppati riproducono il comportamento dei sistemi complessi e valutare la loro risposta alle perturbazioni, anche quando dati quantitativi sono assenti o alcune componenti del sistema non sono completamente caratterizzate. Questi framework potrebbero trovare applicazione in diversi campi di ricerca, comprese, tra gli altri, la biologia, la medicina e la farmacologia, che spesso si trovano ad affrontare sfide di questo genere.
Fuzzy logic for the modeling and simulation of complex systems
SPOLAOR, SIMONE
2020
Abstract
Complex systems typically display emergent dynamic behaviors, which cannot be outlined by a mere description of their constituting parts. Such systems are ubiquitous in natural and artificial settings, ranging from biology and physics, to engineering and economics, and spark the interest of many scientists belonging to different research fields. To gain better insights on their inner working, complex systems are studied by means of computational methods, which allow to model and reproduce their behavior in silico. As a consequence, several mathematical formalisms were developed in the last decades to model complex systems, capture different features of their dynamic behavior, and leverage any available quantitative or qualitative data about their components. The aim of this thesis is to show how fuzzy logic can be exploited to overcome some of the limitations that still affect the modeling of complex systems, namely: predict emergent dynamic behaviors even when there is a lack of precise quantitative information; deal with the presence of heterogeneous systems components, spanning different levels of temporal, spatial of functional organization; bridge the gap between quantitative and qualitative modeling, in order to define hybrid models that can simultaneously exploit the peculiar advantages provided by both approaches. The novel modeling frameworks presented in this thesis based on fuzzy logic have been employed to analyze the behavior of two real world systems, in the context of cellular biology. The results show that the developed frameworks correctly reproduce the behavior of complex systems and assess their response to perturbations, even when quantitative information is missing or some system components are not fully characterized. These frameworks could find applications in several fields, including, but not limited to, biology, medicine and pharmacology, which often face such challenges.File | Dimensione | Formato | |
---|---|---|---|
phd_unimib_820224.pdf
accesso aperto
Dimensione
9.38 MB
Formato
Adobe PDF
|
9.38 MB | Adobe PDF | Visualizza/Apri |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/72994
URN:NBN:IT:UNIMIB-72994