Prof. Itzhak BENENSON, Geography, University of Tel-Aviv, Israel. Geosimulation treats the city as a creature, the complexity of which is above the complexity of physical and chemical systems, but below the complexity of a human self. It thus assumes that there is no need to directly account for real complexity of urban inanimate and animate objects when formalizing urban phenomena. Instead, we could succeed with the agents, which exhibit simple human-like activities that drive the city and its dynamics. The decade of accumulating of the high-resolution GIS, Remote Sensing, population and movement data resulted in cardinal change in the data availability. New sets of data feed Geosimulation models with the adequate behavioral rules and likelihood estimates of parameters, thus bringing us closer to the ultimate goal of Geosimulation - spatially explicit dynamic modeling of urban phenomena. However, these rules necessarily reflect the bounded rationality, i.e., essential uncertainty, of the human behavior. Based on several examples related to the fields of urban and regional planning, land-use dynamics, residential dynamics and urban traffic, I analyze the reasons why some of the Geosimulation models succeed, while some of them do not. I further suggest employing Geosimulation as a tool for understanding inherent uncertainty of urban dynamics and, in this way, for adequate estimating of our ability to predict complex urban phenomena.