Urban Land Use Compatibilities Evaluation Using An Artificial Neural Network Approximator In Micro-Scale

Since the world is rapidly urbanized and urban areas are dealing with enormous changes in utilizing the land, a seamless land use control planning in micro-scale is crucial to prevent incongruous land developement. In this regard, this paper focused on determining the compatibility of parcels in 12th region of Tehran municipality, i.e., one of the oldest and the most populated regions of Tehran, using an artificial neural network approximator. Land use types were divided into 6 types, including: residential, medical, park, religious, educational, and industrial. For each parcel, an effective neighbourhood considered and the percent of whole neighborhood area that occupied by each land use type was calculated. 30 experts point of views were asked on this data to define the degree of compatibility for 250 parcel samples by 1,2,3,4, and 5, corresponded with high compatibility, moderate compatibility, Neutral compatibility, moderate incompatibility and high incompatibility, respectively. Then this data has been formed and fed to the designed multilayer neural network approximator with 10 neurons in one hidden layer and Levenberg-Marquardt backpropagation training algorithm. According to the resulted compatibility map, just 3 percent of this region determined to have a high compatibility with the urban area, where 71%, 8% ,7% and 11% of this region achieved the degree of compatibility of 2,3,4 and 5, respectively. Most of the parcels that fall into the entirely residential area had a moderate compatibility, and higher degree of incompatibility was resulted by the industrial and educational land uses. The results showed that the current situation of land uses is not in accordance with living standards in urban environment. Negligible amount of parcels with high degree of compatibility shows the need for long-time plans to mitigate related land concerns in this region. In this regard, some unpleasent landuses such as industrial land uses should move to vicinty of the city and replace with facilities, sport, and green spaces to mitigate harmful effects on the living population.

کلیدواژه ها:

Land Use Compatibility ، Artificial Neural Network ، Levenberg-Marquardt Back-Propagation ، Micro-Scale