人工神经网络在中国区域可持续发展研究中的应用

Artificial Neural Networks in Regional Sustainable Development Study in China

人工神经网络由于本身所具有的独特优势,使得自然要素环境质量评价更加科学和客观,也使得区域可持续发展的高精度趋势预测和预警成为可能,但在资源可持续利用评价和区域可持续发展综合评价方面的应用尚未取得预期进展.研究表明,人工神经网络在区域可持续发展中的应用不断拓展,但发展态势并不均衡;BP网络虽然应用广泛,但研究中常忽略训练样本与网络结构相互牵制的关系;国内研究多重训练而轻检测,且模型的作用经常被夸大.人工神经网络所推动的区域可持续发展研究新格局有望通过与传统模型、空间信息技术等的交叉和融合而得以进一步的强化和发展.

Artificial neural networks not only make the environmental quality evaluation of the natural elements more scientific and objective,but also make it possible to realize high-precision trends forecasting and early warning in sustainable development of resources and environment for their unique advantages.At the same time,applications in evaluation on sustainable utilization of resources,comprehensive evaluation,forecasting and early warning on the whole regional sustainable development system have not yet made the expected progress.Studies showed that,primarily,applications of artificial neural networks in regional sustainable development are expanding,while the situation and trend of development are not balanced;secondly,back propagation（BP） network has been widely used,while the relationship between the training samples and network structure is often ignored;finally,the training process is emphasized,while the detecting one is often neglected in China,and the role of models has often been exaggerated.The new pattern of regional sustainable development study promoted by artificial neural networks is expected to be further enhanced and developed with the intercross and integration of traditional models,spatial information technology and expert system.