In the 20 th century, the environment of Central Europe was shaped by an extensive growth of urban areas leading to sprawling agglomerations. While the cities' morphological growth is still proceeding, a second major trend is emerging nowadays: urban decline. Accordingly, the polycentric agglomeration of the Ruhr (North Rhine-Westphalia, Germany) simultaneously faces a demographic decline and a physical extension. The modeling of both trends is essential in order to estimate their social and ecological impacts. Among urban land-use models are artificial intelligence techniques like cellular automata (CA) and multi-agent systems (MAS). While CA focus on discrete spatial entities, MAS are well-suited to capture individual decision making. This study presents an approach dealing with the integration of both complementary methods: the coupling of the MAS ReHoSh (Residential Mobility and the Housing Market of Shrinking City Systems) and the CA SLEUTH. SLEUTH is one of the best-assessed spatiallyexplicit urban growth models applied in numerous studies all over the world. Here, the CA will be guided by support vector machines in order to enhance its modeling performance. ReHoSh is a newly implemented MAS catching the interactions between stakeholders of housing markets and the development of potential residential areas in a declining urban environment. The concept of semi-explicit urban weights is introduced transferring the probable dwelling demand as results of individual decision making into the cellular environment. The CA-MAS combination is calibrated in order to mine the urban future of the Ruhr. Beside a "business as usual"-scenario, two further scenarios of changing housing preferences are simulated for 2025. They reflect the dissemination of sustainable thinking among stakeholders and the steady dream of owning a house in sub-and exurban areas. The created total probability maps clearly influence the future rates of SLEUTH. The CA is successfully provided with scenarios resulting in different extents of the Ruhr's urban area for the year 2025: 136,007 ha ("business as usual"), 134,285 ha ("sustainable thinking"), and 140,141 ha ("dream of owning a house"). The spatial impacts are visualized with the concept of urban DNA and a digital petri dish. Here, it becomes obvious that a sprawled pattern of the cities of the Ruhr is just prevented in the scenario "sustainable thinking". Zusammenfassung: Die Zunahme von Siedlungs-und Verkehrsflächen und die damit verbundenen ökologischen Probleme wie Landschaftszerschneidung oder Flächenversiegelung stellen in Deutschland eine noch nicht gelöste Herausforderung dar. Nach neueren Erkenntnissen geht die Flächeninanspruchnahme zwar zurück, doch ist die Erreichung des 30-Hektar-Ziels nach wie vor in weiter Ferne. Gerade auch in schrumpfenden Regionen ist eine ungebremste Ausweitung von Siedlungs-und Verkehrsflächen zu beobachten. Die Gründe für diese Entwicklung lassen sich unter anderem in der angebotsseitigen Bereitstellung von Wohn-und Gewerbeflächen im Zuge einer wachsende...

No abstract