Search device, searching method, and plasma processing apparatus

1 . A method of searching a semiconductor processing device which performs predetermined semiconductor manufacturing processing for an input parameter value of each control parameter of the semiconductor processing device which corresponds to a goal output parameter value to be given to the semiconductor processing device by a processor, the method comprising:

learning a prediction model, by the processor, from learning data which is sets containing a plurality of the input parameter value of the semiconductor processing device and output parameter values which are results of the semiconductor manufacturing processing that the semiconductor processing device performs on the basis of the input parameter values;

setting a target output parameter value, by the processor, by interpolating values between the goal output parameter value and an output parameter value which is the closest to the goal output parameter in the output parameter values in the learning data;

estimating input parameter values, by the processor, which correspond to the goal output parameter value and the target output parameter values by using the prediction model;

determining, by the processor, whether the output parameter value, which is the result of the semiconductor manufacturing processing that the semiconductor processing device performs on the basis of the estimated input parameter value, converges to the goal output parameter value;

in a case where the processor determines non-convergence, updating the prediction model, by the processor, by adding a set of the input parameter value that the processor estimates, and the output parameter value which is the result of the semiconductor manufacturing processing that the semiconductor processing device performs, on the basis of the estimated input parameter value to the learning data, and resetting the target output parameter value, to cause the processor to search for of an input parameter value which corresponds to the goal output parameter value;

in a case where the processor determines convergence, outputting the input parameter value with which the processor determines the convergence as the input parameter value which corresponds to the goal output parameter value, and transferring the input parameter value estimated by the processor which corresponds to the goal output parameter value as an optimal processing condition to the semiconductor processing apparatus via an output device; and

processing a semiconductor device, by the semiconductor processing device, based on the optimal processing condition received from the processor, or processing the semiconductor device, by the semiconductor processing device or a synthesis device which synthesizes materials based on the optimal processing condition received from the processor.

2 . The method according to claim 1 ,

wherein the processor is further configured to

set a plurality of initial input parameter values in such a manner that each input parameter used in the prediction model includes three level values on the basis of settable ranges of input parameters of the semiconductor processing device; and

learning the prediction model by using a set of the initial input parameter values and an initial output parameter value which is a result of semiconductor device processing that the semiconductor processing device performs on the basis of the initial input parameter value as initial learning data.