Genetic algorithms solution to automated zone design based on urban population map

One of the most crucial needs for every researcher who use statistical data at different scales specific at urban scale is designing and defining areal unit of analyses. Some experts use official zoning system and some other use a purpose-based areal units and new zoning system regarding their need and goals of the study. In a city or region, administrative zones are usually designed based on political and administrative intentions. Urban researchers study the city at different levels from local (neighborhood) to the global scale of the city. Here the modifiable areal unit problem shows up. One of the best functional solutions for MAUP problem is using Genetic Algorithm techniques Genetic algorithms (GA) are subclasses of Evolutionary Computing. This method can counter the effects of MAUP on spatial based statistical indexes and results. In this paper GA has been used for zone design and portioning city into proper districts for produced (objective) accessibility raster map in Tehran which is a practical real world problem. The GA accounts for essential characteristics population equality, contiguity, geographical compactness. The result shows significant improvements in the matters of easiness of application and fastness for Automated Zone Design (AZD) in real world problems also add some general concepts to show the benefits of our work for urban planners and spatial data researchers and analyzers who face the MAUP.