System and method for predicting forest pest disaster

Abstract

The invention discloses a system and a method for predicting a forest pest disaster. The system mainly comprises a forest pest disaster prediction model module, wherein the forest pest disaster prediction model module comprises a geographical cellular automata model, an artificial neural network model and a multi-agent model; each cell in the geographical cellular automata model represents a geographical area and has a cell state which is used for representing disaster degree and one or more cell attributes which are used for representing disaster influence factors; a state conversion rule is acquired by training the artificial neural network model; the input of the artificial neural network model is the disaster influence factors, and the output of the artificial neural network model is the disaster degree; the multi-agent model comprises a human activity influence agent which is used for representing influence of human activity on the cell state, and a pest population change agent which is used for representing the dynamic evolution process of a pest population; and analysis results of the agents are integrated with the current cell state to obtain the updated current cell state.

Description

technical field [0001] The invention relates to the field of forestry harmful biological disaster prediction, in particular to a forestry harmful biological disaster prediction system and method. Background technique [0002] Biological disasters, including pests, diseases, rats (rabbits), harmful plants and other disasters, have caused huge impacts and losses on the sustainable management and sustainable development of forestry resources. Therefore, a lot of research has been done on the prediction of forest pest disasters at home and abroad, and many prediction methods and prediction models have been proposed, including discriminant analysis models, principal component analysis forecasting methods, time series models, statistical models, gray predictions, and Markov chains. , spatio-temporal regression forecasting methods and so on. [0003] The current forestry pest prediction models and methods have the following main problems: [0004] At present, most predictions are...

AI Technical Summary

Problems solved by technology

There are too many types of agents in the model, and the whole system is very complex, resulting in very slow calculation speed, and too few will not reflect the complexity of the system

However, the factors that affect the occurrence and spread of forest pests and diseases are complex, in addition to natural factors, they are also affected by unpredictable social factors

For example, various natural enemies of the host, human activities (felling of diseased and infected wood, transportation, etc.), road network and village distribution, etc., if considering habitat, human activities, natural enemies and other uncertain factors, the MAS system will be very complicated , causing the forecasting process to be too long, or even paralyzed

Images