Control device for generating a control deviation via machine learning

1. A feedback control device that takes information regarding a control deviation between a measured value and a desired value of an object as input, and outputs a feedback manipulated variable for the object, comprising:

a processor configured to:

obtain information regarding the control deviation as input and generate a first manipulated variable for the object;

input information regarding the control deviation to machine learning and to generate a second manipulated variable based on a parameter determined by the machine learning;

set a limit to the second manipulated variable; and

add the first manipulated variable and the limited second manipulated variable into a feedback manipulated variable; and

control the object based on the feedback manipulated variable,

wherein the processor is further configured to function as a filter for stopping a certain band of the control deviation, and

a signal from the filter is input for the machine learning.

2. The feedback control device according to claim 1 , wherein the second manipulated variable is limited within a predetermined range by the processor.

3. The feedback control device according to claim 2 , wherein the parameter is machine-learned in a state where the range of the second manipulated variable that may be output is limited by the processor.

4. The feedback control device according to claim 2 , wherein the range of the second manipulated variable is smaller than a range of the first manipulated variable.

5. The feedback control device according to claim 1 , wherein:

the processor further functions as a plurality of the filters having different stopbands,

wherein the plurality of filters are switched and used.

6. The feedback control device according to claim 5 , wherein the plurality of filters comprises:

a first filter and a second filter, and

the parameter includes

a first parameter determined by machine learning using a deviation in a first band as input in a case where the first filter is used, and

a second parameter determined by machine learning using a deviation in a second band different from the first band as input in a case where the second filter is used.

7. A system comprising the feedback control device according to claim 1 , wherein the processor is further configured to function as:

a driving unit configured to apply a change to the controlled object; and

a measurement unit configured to measure the controlled object,

wherein the feedback control device takes information regarding a control deviation between a measured value, measured by the measurement unit, and a desired value of the controlled object as input, and outputs a manipulated variable for the drive unit.

8. A lithography apparatus comprising:

the feedback control device according to claim 1 ; and

a forming unit that forms a pattern using an object controlled by the feedback control device.

9. An article manufacturing method comprising:

a step of forming a pattern on a substrate using the lithography apparatus according to claim 8 ;

a step of processing the substrate on which the pattern is formed; and

manufacturing an article from the processed substrate.

10. A measurement apparatus comprising:

the feedback control device according to claim 1 ; and

a measurement unit that measures an object controlled by the feedback control device.

11. A processing apparatus comprising:

the feedback control device according to claim 1 ; and

a processing unit that processes an object controlled by the feedback control device.

12. A planarizing apparatus comprising:

the feedback control device according to claim 1 ; and

a planarizing unit that planarizes a composition using an object controlled by the feedback control device.