Studies on Intelligent Decision and Support System for Forest Fire Suppression

Forest fire spreads freely in the opening systems. With the effect of the terrain, the types of fuel and the impact of weather conditions, the process of combustion is complex and changeable. Each forest fire has different characteristics; besides the knowledge and technology that fire fighters possess also vary. Because fire is complex and changeable, it’s difficult to produce a reasonable project only by the commander’s experience.The paper marks out the whole structure of the intelligent decision and support system for forest fire suppression, and consummates the whole layout. The intelligent decision and support system for forest fire suppression consists of five core modules which use parallel work style. These are fire information collection, fire action prediction, fire suppression intelligent decision, user interface and database. Forecast module has used the knowledge base, the global database, the inference, the reasoning machine and the model library circulation working pattern, and through interface realized coordination multi-tasking with the intelligent decision and support system for forest fire suppression. Produced the fire spread simple configuration with superpose vector take Wang Zhengfei forest fire spread model as the foundation; Produced the model to changed-length of stride forecast method; produced the multi- models forecast intelligence choice method in view of Wang Zhengfei model and GM(1,1) characteristic; in forecast process, constructing the necessary information with specificity and handy principle for predicting and calculation.After investigating the decision method based on resource competition, the paper adopts the method which enables the system to decide the optimal fire suppression method when the fire suppression resources are limited. At the same time considering of fire suppression time, fire area, fire loss and so on, the paper introduces a series of estimation system. Optimized the traditional fire-fighting way; and through analysing the landform of forest fire, establish a decision-making function judging the dangerous geographical environment which includes the steep slope gradient, single-mouth valley, saddle-shaped valley, by calculating grade. And the function makes the decision-project more scientific and reasonable.On the basis of analysis to a large number of historical cases of forest fires suppression, through researching the knowledge expression and structure of knowledge database, and in accordance with the characteristics of intelligent decision and support system for forest fire suppression, using the knowledge expression based on“rule skeleton + rule body”, established the knowledge database. And researched the automatic learning algorithm, developed the learning machine that can gain the knowledge automatically, which based on the natural language matched with prototype. In this study, the specialized dictionary which supported the understanding of the natural language has been established, and learning machine’s learning efficiency has been improved by the words’ Second localization.