Reloading scheme search optimization method

The invention relates to a refueling and loading scheme search optimization method. The method comprises the following steps: S1, generating a plurality of common new component layouts according to a plurality of historical refueling schemes; obtaining old component information in the historical refueling scheme, and generating an old component database; s2, selecting old components from an old component database according to the common new component layout, randomly loading the old components into the common new component layout to generate K loading schemes, and selecting K1 loading schemes from the K loading schemes to form training samples; wherein K1lt; k; s3, training by using the training sample to generate a prediction model library meeting a precision requirement; the prediction model library contains a plurality of key parameter prediction models; and S4, selecting a key parameter prediction model corresponding to the refueling notification information from the prediction model library according to