BIG DATA SYSTEM RESOURCE CONFIGURATION PARAMETER TUNING METHOD BASED ON GENERATIVE ADVERSARIAL NETWORK

The present disclosure relates to a big data system resource configuration parameter tuning method based on a generative adversarial network. The method comprises the following steps: S100, constructing an evaluated configuration parameter set, and recording optimal configuration parameter values in the evaluated configuration parameter set; S200, constructing a Gaussian process by using the evaluated configuration parameter set, and calculating a mean value and a variance; S300, constructing a candidate configuration parameter set, wherein the set at least comprises N1 groups of configuration parameter values, which are generated by a generative adversarial network simulating the current optimal configuration parameter values; and S400, substituting the variance and the mean value into an acquisition function, finding, from the candidate configuration parameter set and by using the acquisition function, a group of configuration parameter values which possibly have the optimal performance, and taking same as