Landslide displacement prediction method based on intuitionistic fuzzy memetic PSO-LSTM

Abstract

The invention discloses a landslide displacement prediction method based on an intuitionistic fuzzy memetic PSO-LSTM model. The method comprises the following steps: 1, establishing an initial LSTM model; 2, initializing a sub-population; 3, executing a collision rebound operator when the aggregation degree of the exploring sub-populations reaches a threshold value, otherwise, executing the step 4; 4, updating the exploration sub-population, calculating the fitness and updating the optimal solution; 5, calculating the Lamarch factors of the particles in the development sub-population; 6, updating the speed and position of the particles, calculating the fitness of the particles, and updating the optimal solution; 7, judging whether the development sub-populations converge or not, and if yes, selecting an optimal target area; 8, if the end condition is not met, returning to the step 3; 9, explaining the optimal solution as LSTM; and training the LSTM model by using the training sample. According to the invention, the heuristic optimization algorithm is combined with the intuitionistic fuzzy set, the LSTM prediction model is optimized and is more suitable for the landslide prediction problem, and the prediction precision of the prediction model on the lively problem is improved.

Description

technical field [0001] The invention belongs to the technical field of landslide control, and in particular relates to a landslide displacement prediction method based on an intuitionistic fuzzy dense parent PSO-LSTM model. Background technique [0002] At present, there are many landslide displacement prediction methods in landslide control technology, and Recurrent Neural Network (RNN) is one of them. In this method, Long Short-term Memory Networks (LSTM) is a commonly used and powerful A deep learning model with good time series forecasting ability. By associating with the previous unit on the sequence, RNN enables the previous unit to affect the calculation of this unit, so it can process sequence information efficiently and orderly. LSTM is further improved on the basis of RNN. By adding forgetting gates, input gates, and output gates, it can effectively avoid the short-term and long-term dependence problems of RNN, making it more suitable for processing time series. ...

AI Technical Summary

Problems solved by technology

[0003] However, there are the following objective problems in landslide displacement prediction: the acquisition of landslide data is usually easily affected by objective factors such as environmental factors (groundwater level, rainfall, human activities, earthquakes), equipment factors (equipment accuracy, equipment damage), etc., resulting in The collected sample data often has problems of large scale, nonlinearity, high information redundancy, ambiguity, uncertainty, and high noise. These problems greatly affect the direct application ability of LSTM on this problem.

[0004] In summary, the existing landslide displacement prediction technology has the problems of mismatching artificially designed prediction models and difficulty in improving the accuracy due to factors such as high-dimensional, fuzzy, abstract, and uncertain data characteristics.

Images