US 11,112,775 B2Granted

System and method of determining processing condition

AssigneeHitachi, Ltd.

InventorsHyakka Nakada, Takeshi Ohmori, Tatehito Usui, Masaru Kurihara

Life of this patent

Estimating the term…

Loading patent…

Abstract

A system for determining a processing procedure including a plurality of processes for controlling an object, the system includes a learning unit for performing a learning process for determining a processing condition of each of a plurality of processes, and the learning unit acquires a physical quantity correlating with a state of the object on which a process has been performed under a predetermined processing condition, from a device for controlling the object on the basis of the processing procedure, calculates a pseudo state corresponding to the state of the object on the basis of the physical quantity, performs a learning process using a value function, and determines a processing condition of each of the plurality of processes to achieve a target state of the object.

Claims10

Dependent claims

A system for determining a processing procedure including a plurality of processes for controlling an object, the system comprising:

at least one computer including an arithmetic device, a storage device connected to the arithmetic device, and an interface connected to the arithmetic device, wherein
the plurality of processes is each performed on the basis of a preset processing condition, and
the at least one computer is configured
to perform a learning process for determining the preset processing condition of each of the plurality of processes,
to acquire, from a device for controlling the object on the basis of the processing procedure, a physical quantity correlating with a state of the object for which a process included in the processing procedure is performed under a predetermined processing condition to calculate a pseudo state corresponding to the state of the object in reinforcement learning,
to determine a processing condition of each of the plurality of processes for achieving a target state of the object by using a value function for outputting a value indicating an evaluation of the processing condition selected in an appropriate pseudo state,
to acquire the physical quantity to calculate the pseudo state on the basis of the physical quantity,
to select an appropriate processing condition from among the processing conditions of the process performed next, on the basis of the pseudo state and the value function,
to output a request for performing processing including the selected processing condition to the device, and
to acquire the physical quantity from the device having performed a process on the basis of the request for performing processing and to calculate the pseudo state, and
wherein
the object is a semiconductor processing apparatus for manufacturing a semiconductor by processing a sample, and
the processing condition is a condition for controlling the semiconductor processing apparatus.

10 more elements

A method of determining a processing condition set in a plurality of processes included in a processing procedure for controlling an object, the method being performed by a system,

the system including at least one computer having an arithmetic device, a storage device connected to the arithmetic device, and an interface connected to the arithmetic device,
the plurality of processes each performed on the basis of a preset processing condition,
the method of determining the processing condition comprising:
a first step of acquiring a physical quantity correlating with a state of the object for which one process included in the processing procedure has been performed under a predetermined processing condition from a device for controlling the object on the basis of the processing procedure to calculate a pseudo state corresponding to the state of the object in reinforcement learning on the basis of the physical quantity, by using the at least one computer;
and
a second step of determining a processing condition of each of the plurality of processes for achieving a target state of the object by using a value function using the pseudo state and processing condition as variables to output a value indicating an evaluation of the processing condition selected in an appropriate pseudo state, by using the at least one computer,
wherein the second step includes
selecting an appropriate processing condition from among the processing conditions of the process performed next on the basis of the value function and the pseudo state, by using the at least one computer, and
outputting a request for performing processing including the selected processing condition to the device, by using the at least one computer, and
the first step includes acquiring the physical quantity from the device having performed a process on the basis of the request for performing processing, by using the at least one computer, and
wherein
the object is a semiconductor processing apparatus for manufacturing a semiconductor by processing a sample, and
the processing condition is a condition for controlling the semiconductor processing apparatus.

10 more elements

Dependencies are inferred from the claim text.

Facts

Application: 16/430,844Filed 4 Jun 2019
Priority: JP JP2018-1074195 Jun 2018
Published: US 2019/0369605 A15 Dec 2019
Granted: 7 Sep 2021
Term ends: Not estimatedThe source records don’t support an estimate.

Classification10 codes

G05B 19/41805
Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible…:…
G05B 13/0265
Electric: the criterion being a learning criterion
G05B 19/41865
Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible…:…
G05B 2219/32334
Operator till task planning: use of reinforcement learning, agent acts, receives reward