关键词: 商业网点规划, 哈夫模型, 路网距离, 市场服务区

Abstract: Changzhou City，Jiangsu Province is to the south of the Yangtze River，and in the middle between Shanghai and Nanjing City. It is a prosperous city with land area of 4 372.15 square kilometers. In 2012，the area of urban built up is 183.24 square kilometers，The GDP of the city is 396.98 billion RMB. The ratio of the first industry，second industry and third industry is 3.2∶52.9∶43.9. The average personal GDP in Changzhou City is 85 039 RMB，or 13 471 USD，ranked 4th among 13 cities in Jiangsu Province. Changzhou has 5 districts and 2 county-level cities. To promote the sound development of Wujin District of Changzhou City’s business cycles，the paper modified the Huff model to delineate market domains of Changzhou and Wujin business centers. Huff model has been widely applied in urban retail market domain planning since its introduction. Huff model is applied to calibrate interactions between Central Business Districts （CBDs），surrounding enterprises，and consumers. The model shows high accuracy in predicting business cycles’ market shares. Also，when Huff model is applied to delineate market domains for a number of business cycles，the market domains appear as gradational cycles，which are closer to actuality. After 40 years of applications，the interpretability，convenience，and flexibility of Huff models have gained wide accolade from businesses，governmental analysts，and scholars. To determine market domains for business cycles of different scales，the paper suggested multiplying the distance from consumers’ to the secondary business cycles by the distance to the primary business cycles in place of the square of consumers’ distance to one business center. This modified [λ] works better for delimiting market domains for business cycles of different scales. Also，the paper proposed that business attraction should be determined by mid-sized to large retails. Moreover，our research compares straight line distance，network distance，and network time to provide alternative insights for business planning. The conclusions are as follows：（1）the choice of three types of distances （straight line distance，network distance，and network time） has great impacts on results of Huff model. Survey method is optimal to obtain the distance. When data collection is difficult，network distance and travel time are closer to reality；（2） the business attraction should be determined by the areas of mid-sized and large retails. This makes survey easier and works well；（3） the Tajima modification of Huff model is convenient and interpretable，but it cannot be applied to delineating market domains for business cycles of different scales. In such occasions，the modification of [λ] in this paper can be adopted with survey method （if possible）；（4） it is also crucial to reasonably determine region sizes of consumer orientations. Small region sizes generate huge computing load，while overly large region sizes affect the accuracy of Huff model. With those four main points，this paper helps users to figure out a balance between the needs for accuracy and costs of survey and analysis.

Key words: urban commercial network planning, huff Model, network distance, market area