| The location is very important in the operation of General Merchandise Stores. As the retail format of General Merchandise Store has been introduced to China only for a short time, there are only few researches on location of General Merchandise Stores. As a result, being lack of effective theoretical direction on location, some General Merchandise Stores fell into trouble because of located at wrong places. Therefore, providing a scientific location evaluation method for operators of General Merchandise Store in China is of very important significance both theoretical and practical.Based on review and analysis of theories both home and abroad, the thesis pointed out tow problems. Firstly, researches on location of General Merchandise Store in China are not enough, and none of western retailing location theories can guide the location of General Merchandise Store independently, although there are many retail location theories there. Secondly, there are some problems of existing retail location evaluation methods and models– they are inconsistent with conditions in China or there is insurmountable problem the method or model itself. Therefore, it is necessary to introduce new methods of location evaluation. Artificial Neural Network is an interdisciplinary subject, demonstrating its superiority in modeling of complex systems and issues. As one kind of Artificial Neural Network, BP Neural Network has the characteristics of high-speed computing and learning, with the capacity of approximating any nonlinear function theoretically. It is very effective in application of forecasting and evaluating, especially under complex decision-making situation. Therefore, this thesis will introduce BP Neural Network into location evaluation of General Merchandise Store.According to the principle of BP Neural Network, the author proposed the principle of the General Merchandise Store location evaluation method based on BP Neural Network, and the location evaluation system of General Merchandise Store based on combining of western retail location theories, factors affecting retail location improved in related researches and results of in-depth interviews. Then, the author designed the network structure of the location evaluation model of General Merchandise Store, according to the location evaluation system proposed before and BP Neural Network theories. The author gained 23 samples through semi-structured interview, observation and questionnaire investigation. Then, the author put 20 standardized samples into the established network under condition of Matlab6.5 to train the network, until it reached the goal. Finally, the author presented 3 cases, and put these 3 samples into the network to check it. The results showed that the location evaluation model of General Merchandise Store based on BP Neural Network is effective, which suggested that it is of practical use in location evaluation of General Merchandise Store. |
