The land-use simulation procedure is spatially explicit, operating on cells of
100!100 m2 size, wherein conversion probabilities of a cell are computed for different
uses. In a CA model, the state of a cell at time (tC1) is decided by the state of its
neighbouring cells at time (t) according to pre-defined transition rules [15,18]. In the
CA-MCDM approach adopted here, the future state of a cell is determined based on the
transition probability of a cell for a given use. This transition probability in turn is
determined as a function of not only the neighbourhood effects but also other important
determinants of land-use changes such as land characteristics (e.g. soil, topography),
current land-use, proximity factors (such as distance to markets) as well as human factors
including user preferences for evaluating criteria and the user responses to questions
concerning preservation of agricultural land, agricultural intensification and management
practices, and diet preferences. The responses related to intensification of agriculture and
diet preferences change the area requirements for different land-use types and affect landuse
transitions in an indirect way by assigning competitive advantage in the spatial
allocation process to the land-use types that are in greater demand.