Presentation of a model for regionalization of urban areas using AZP algorithm

urban planners in order to detection and resolving of urban problems would need to deep recognition of city and planning for it. Because cities are composed of most widely societies and varies in infinite dimensions and urban planner couldn’t devise and apply same plan for whole city, the city must be divided into spatial units based on social, economic and physical aspects. This segmentation then used for some constraint like specific land use or construction density and so on or for controlling development of city. This kind of segmentation be defined under lots of urban plans like detailed plans, comprehensive plan etc. In the other hand in a city there is lots of organizations like Municipalities, Roads and Urban Development organizations, Tavanir companies, Abfa companies etc. which have made their own segmentations for servicing their customers. All of these configurations have been made for responding specific requirements and receiving some goals and with respect to some approaches but Most of the time these segmentations have functional overlap and In many cases conflict with each other and aren’t efficient enough and cause to destroying sources and making people unsatisfied because they oblige to explore lots of organizations to do their work. In the other hands the number of units and divisions of all Municipals, organizations and agencies isnt fixed and have been increased through the time due to the increasing population and urban area. Such a situation with inconsistence spatial structure and lack of public participation, are some reasons which make urban sustainable management difficult. So the aim of this article is to apply new approaches and methodologies to solving this problems. The AZP is one of the many different methods to do that. AZP is an algorithm which used extensively for delimitation of multidimensional units or detection of spatial scale to study specific relationships in space of a city. In this article we want to apply this model in an urban areas of Iran which weve chosen Zanjan city as a case study. spatial clustering methods (such as AZP) have been developed during many years and they could be used for making regions, zones etc. by aggregating and interchanging basic spatial units in each other with optimization of objective functions. These regionalization algorithms have some basic features like: all of them aggregate basic spatial units into predefined number of regions with optimization of an aggregation function, basic spatial units which assigned to a region must be spatially connected, the maximum number of regions should be one less than basic spatial units, a basic spatial unit could be aggregated to only one region and minimum number of basic spatial units which is should be assigned to a region is one. Another characteristics of them is their supervise capability. That’s because of this you should define relevant variables, number of regions and type of objective function. The AZP algorithm can work with any type of objective function that is sensitive to the aggregation of data for N basic spatial units into M regions for example functions extracted directly from the data (for instance sum of squared deviations from average zone size) or functions that can represent the goodness of a fit of a model applied to data (fit of a linear regression model or the performance of a spatial interaction model). Output regions will not change over the time and would make urban management more sustain and more coordinated so these configuration could be used as basic directorial units of different organizations. Type of research is applied and purpose of us is development of unified multidimensional regionalization which could assist sustainable management of city. We used census block and land use of Zanjan city as basic data. Then we extracted 26 indicator from those data and aggregated them into fishnet with cell size of 300 meter. In the next step for decreasing heterogeneity and discovering general trends we applied principle component analysis (PCA) on our indicators so five PC extracted which control 76 percent of variance. Regionalization have been conducted based on these PSs. Objective functions are an intra-area correlation function and a shape function which optimize output regions. The homogeneity of regions can be evaluated based on using a direct measure of within-area homogeneity, the intra-area correlation (IAC). We measured shape compactness by comprising the squared perimeter with the area of each output area. Finally after running algorithm and exporting results we need to test our output. Moran's I statistic is a measure of spatial autocorrelation in which Negative (positive) values indicate negative (positive) spatial autocorrelation and it's Values range from −1 (indicating perfect dispersion) to +1 (perfect correlation) and A zero value indicates a random spatial pattern. we used 13 factor for validation of our work. Resulted regionalization have been compared with regionalization of detailed plan of Zanjan city for validation of AZP algorithm based on Moran's I statistics. The Moran's I showed us in contrary with detailed plan regionalization which have clustered spatial pattern, AZP algorithm could create regions with random spatial pattern which indicate to homogenous regionalization. this research showed that unknown pattern should be recognized by new methodologies so urban planners could do planning more efficiently. We shouldn’t limited our research to earlier constructed regions because every regionalization have been made for specific purpose and certainly would affect the result of analysis so at first step there is a need to make our objective-related regions. This regions then could be a very statistically meaning spatial units which is free from MAUP and very reliable.