Oil reservoir numerical simulation parameter inversion and optimization method based on intelligent agent

The invention provides an intelligent agent-based numerical reservoir simulation parameter inversion and optimization method, and relates to the field of oil and gas reservoir development, and the method comprises the steps: building an intelligent agent automatic simulation system, and building automatic program control numerical reservoir simulation software; performing intelligent agent data processing, and constructing an intelligent agent data set; establishing an intelligent agent prediction model, selecting a learning model, training the learning model by using an intelligent agent data set, and performing hyper-parameter optimization on the learning model with the optimal performance; performing intelligent agent inversion parameter optimization, selecting an inversion parameter optimization algorithm, establishing a target function, setting the population size of the inversion parameter optimization algorithm, and outputting an optimal inversion parameter by the intelligent agent inversion parameter