1、基于 Web 的农业智能系统开发平台研究- 1 摘 要从农业现代化走向农业信息化，是农业发展的必然趋势。农业智能系统作为农业信息化的重要内容，目前在我国还存在网络化、智能化程度不高等问题。本文在分析研究国内外农业智能系统的基础上，应用现代信息技术和人工智能技术，构建了基于 Web 的农业智能系统开发平台。本文针对农业知识种类繁多、关联性强的特点，提出了一种适用于农业智能系统的面向对象知识表示方法。该知识表示方法便于农业知识的管理、检索和重用，有效地提高了推理机的推理效率。并在对传统优化算法和一般人工免疫优化算法的分析和研究基础上，从免疫系统中免疫响应机制出发，根据免疫响应亲和力成熟的过程，提出
2、了一种改进的人工免疫优化算法。这种算法能够在保持全局搜索能力的同时有效地进行局部搜索，一定程度上解决了进化计算通常面临的早熟收敛问题。将该算法应用于施肥子系统中，取得了较好的效果。本平台采用 J2EE 框架、面向对象思想、人工智能等技术进行构建，实现了基于面向对象知识表示的远程推理。平台具有良好的动态可重构性、开放性、可扩展性和可跨操作系统等特点。基于该平台已开发了包括番茄、辣椒、甜玉米、茄子以及白菜在内的五种无公害蔬菜生产远程咨询系统，并投入实际应用。关键词: 农业智能系统开发平台；人工免疫优化算法；面向对象；知识表示；Java基于 Web 的农业智能系统开发平台研究- 2 Abstract
3、Agriculture develops from industrialization to information times is an inevitable trend. Agricultural Intelligence System (AIS) plays an important role in the process of agricultural information development, however, the AIS in China has some problems such as the low development of network and less
4、intellectualized. Based on the studies of domestic and foreign agricultural intelligence system，the author established an Agricultural Intelligence System (AIS) platform based on web by applying the modern information technology and artificial intelligence technology.The article utilizes the agricul
5、tural knowledge characteristic that has many types and strong connection to proposed one kind expression method. This expression method is suitable for the agricultural intelligence system object-oriented knowledge. The method is advantageous for the agricultural knowledge management, retrieves and
6、entrusts, effectively enhanced the inference efficiency of inference machine. Based on the analyses and research of traditional optimization algorithm and general artificial immunization optimization algorithm, Embarks from the immunity response mechanism of the immunity system, according to the imm
7、unity response affinity mature process, the article proposed one kind of improvement artificial immunization optimization algorithm. This algorithm can maintain the completely search ability while do the partial search effectively, in the certain degree solved the problem of early-convergence that e
8、volutionary computation usually faces. The fertilizer subsystem has made better result after applying this algorithm to it.To construct this platform, it adopts J2EE frame, object-oriented thoughts, artificial intelligence, the author has achieved long-distance inference based on the object-oriented
9、 knowledge expression. The platform has characteristics such as opening, expansibility, operating, and the good revaluation etc. This platform is utilized at five kinds of harmless vegetables production long-distance consultation system. The vegetables are tomato, pepper, sweet corn, eggplant and cabbage. Keywords: AIS developing platform, artificial immunity optimizes algorithms, object-oriented，knowledge representation, Java
