The problem of modeling cognitive systems has been recently considered under new point of views. The attention is no more mainly focused on the simulation of rational behaviors, but also on the importance of the individual free-will and of the apparent "irrational decisions". A possible approach for a mathematical modeling is to introduce an internal dynamical structure for the "elementary particles" that allows to define an individual cognitive dynamics at the base of the decision mechanisms. Our assumptions imply the existence of an utility potential which is the result of an evolution process, where some behavioral strategies have been selected to perform efficiently in presence of different environmental conditions. The utility potential defines a cognitive dynamics as a consequence of information based interactions and the choice of a strategy depends on the individual internal cognitive state. This is the definition of an "automata gas model", whose aim is to extend some statistical physics results to cognitive systems. We shall discuss the existence of self-organized dynamical states in an automata gas, that are the result of a cognitive behavior. We show as the introduction of a cognitive internal dynamics allows to simulate the individual rationality and implies the existence of critical conditions at which the system performs macroscopic phase transitions. The application to the pedestrian mobility modeling problem is explicitly considered and we discuss the possibility of a comparison with experimental observations.
Cognitive dynamics in a automata gas
BAZZANI, ARMANDO;GIORGINI, BRUNO;RAMBALDI, SANDRO
2009
Abstract
The problem of modeling cognitive systems has been recently considered under new point of views. The attention is no more mainly focused on the simulation of rational behaviors, but also on the importance of the individual free-will and of the apparent "irrational decisions". A possible approach for a mathematical modeling is to introduce an internal dynamical structure for the "elementary particles" that allows to define an individual cognitive dynamics at the base of the decision mechanisms. Our assumptions imply the existence of an utility potential which is the result of an evolution process, where some behavioral strategies have been selected to perform efficiently in presence of different environmental conditions. The utility potential defines a cognitive dynamics as a consequence of information based interactions and the choice of a strategy depends on the individual internal cognitive state. This is the definition of an "automata gas model", whose aim is to extend some statistical physics results to cognitive systems. We shall discuss the existence of self-organized dynamical states in an automata gas, that are the result of a cognitive behavior. We show as the introduction of a cognitive internal dynamics allows to simulate the individual rationality and implies the existence of critical conditions at which the system performs macroscopic phase transitions. The application to the pedestrian mobility modeling problem is explicitly considered and we discuss the possibility of a comparison with experimental observations.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
