| Crop pest control intelligent decision is one of important areas of intelligent decision-making theory and methods’ applications, is an important research content of digital agriculture, which explores the use of space-time correlation method to reveal the relationship between the occurrence of pests and diseases and environmental inducing factors; takes advantage of Modeling and simulation technology, forecast the occurrence, development and trend of disease; uses multi-objective planning or case-based reasoning technology, conduct intelligent and efficient diagnosis and decision-making pest control measures.Crop pests is a complex natural phenomena. In terms of space, it is entirely influenced, but also it has a regional impacted feature. In terms of time it has performed stability of disorder, but has the existence of an orderly and cyclical regularity, with inhomogeneity, difference, diversity, sudden, random, predictability and regularity of complex characteristics. It is because of the complexity of pest disasters, making very difficult for the diagnosis of pests, disease occurrence trend forecasting, prevention methods decision-making. Relationship between decision making factors is not clear, and sometimes decision-making information is incomplete, these owe to the bad structure of decision-making problems.Current expert systems in home and abroad have been related to a certain kind of specific diseases, using traditional knowledge representation rule-based approach, has the "knowledge acquisition" bottleneck and difficulties in knowledge base maintenance and other defects. Especially for the existing pests expert system, the core component system inference engine, usually uses a single case-based reasoning, or reasoning rule-based reasoning, so that its reasoning efficiency is usually low, and this is one of the root causes resulting in pest forecasting systems existing unobvious performance. This paper focuses on knowledge sharing and pest diagnosis, forecasting and decision prevention study, key issues, mainly to complete the work in the following aspects:(1) Crop pests ontology integration. For multi-source information for Crop pests, isomerization and other uncertainties, this paper build a model to study the ontology of Crop pests and diseases, XML expression ontology, ontology mapping and ontology evolution for consistency checking and other issues.(2)Knowledge representation of Crop pests ontology case. First, for the characteristics of the case against pests and diseases, defining Fuzzy ontology language SWR & Fuzzy L (Fuzzy Semantic Web Rule Language), to achieve pest knowledge representation of uncertain cases, make establishment of an ontology knowledge model case of Crop pests of case-based reasoning, to design a structured case base XML/SWR & Fuzzy L mapping algorithm to achieve semi-automatically acquiring of a case.(3) Case-based reasoning retrieval. For key steps of case-based reasoning this paper conduct retrieval research, including cases feature recognition, feature attribute weights for determining the case, the case features attribute reduction and case similarity calculation methods. Attribute reduction method based on A-SVM is proposed on case characteristics, Pest case similarity algorithm based on core matrix iterative algorithm pest’s case is proposed too, and this paper gives reasoning search strategy on case ontology-based and conducts experimental verification.(4) Case-Based reasoning decision pest prototype system. Based on the above findings, combined with Crop pests forecasting workflows, it establishes pests decision model based on ontology and case-based reasoning, do research on case-based reasoning on ontology, makes prototype system of case-based reasoning for pest decision. |
PDF Full Text Request