Automated design of gradual zone systems

摘要: The appropriate resolution of a zone system is key to the development any transport model, as well other spatial analyses. number and shape zones directly impacts effectiveness further modeling steps, with trade-off between computation time model accuracy being particularly important consideration. Currently, systems are often designed by hand. gradual rasterization zoning algorithm produces good empirical results computationally generating raster cells varying area, but similar population employment. We address several limitations original in this paper. Firstly, allocation employment weighted land use instead area percentage. Secondly, extended respect municipal delineations. Raster split along these delineations, undesirably small created process re-merged neighbors same municipality. Aligning generated municipalities simplifies improves disaggregation socioeconomic data. An iterative has been developed automatically determine threshold that desired size. Using algorithm, for Munich metropolitan region. Only 13 iterations were needed converge within 5% target 5000 delineated zones. improved maintains advantages adds improvements useful when creating system.