Single well production dynamic prediction method based on long and short term memory deep neural network

The invention discloses a single well production dynamic prediction method based on a long and short term memory deep neural network. The method comprises the following steps: acquiring an input historical data set X and an output historical data set Y; respectively extracting an input training set X1 and an input test set X2 from the input historical data set X, and obtaining a production dynamic prediction model and an output test prediction set Y2 by applying a long-short term memory deep neural network; screening out a production dynamic prediction model meeting an expected effect through a root-mean-square error function; obtaining single-well production parameters and single-well production parameter constraint conditions, and generating a plurality of injection-production schemes by applying an optimization algorithm; and for each injection-production scheme, acquiring historical production dynamic data corresponding to each injection-production scheme, and sequentially applying the production dynamic