In spatial analysis, it is often necessary to use aggregated data, either because the data is available only for predetermined spatial units, or to make the problem manageable. Zoning is a process where M input zones are grouped into N output zones so that various constraints are satisfied (for example, internal connectivity of the resulting zones, N is less than M). For the purpose of the zoning problem, the exact configuration of the boundaries of the input zones is not important. One possible abstraction is to represent the input zones as nodes of a graph, where each node in the graph represents a zone and each link represents the adjacency relation between zones. Using the graph representation as a convenient framework for solving the zoning problem, a zoning algorithm based on a modified graph partitioning problem (GPP) is proposed. (A)
Samenvatting