The modeling of a city's evolution is one of the most exciting fields in urban studies research. The interest of such a model lies more in the past than the future: understanding how a city was made rather as opposed to how it might evolve. Such a model can necessitate expertise from numerous fields. From urban network analysis to spatial organisation and social patterning, modeling is a matter of choice. Our model simulates a set of alternative planning scenarios and urban development options. City growth reveals important issues to be taken into account in the city planning process and in territorial diagnoses, and analysis of potential alternatives is an essential step for both technicians and policy makers. Using a spatio-dynamic urban evolution model, we discuss the implementation of agent- based modeling in an open-source geographical information system (GIS). Our model simulates the evolution of urban structures as a function of different scenarios at the city block scale in Nantes-Métropole, France.