Device and method for changing a technical system or a technical process

A method for analyzing a technical system that includes a first function and a second function, or for analyzing a technical process that includes a first function and a second function, the method comprising the following steps:

• providing for the first function a first model, and providing for the second function, a second model is provided, the first model modeling a functional relationship of an input variable and an output variable of the first function and an influence of a first disturbance variable as a function of a first parameterization and as a function of a predefined learned distribution over the first parameterization, the second model modeling a functional relationship of an input variable and an output variable of the second function and an influence of a second disturbance variable as a function of a second parameterization and as a function of a predefined learned distribution over the second parameterization;
• providing an observation that includes input variables and output variables of the first function and of the second function;
• determining a prediction regarding an effect of:
• i) changing the first function under an influence of the first disturbance variable, and/or ii) changing the second function under an influence of the second disturbance variable;
• and
• selecting from the first function and the second function that function whose change has a higher likelihood of achieving a desired effect, compared to changing the other of the first function and the second function.

A device for analyzing a technical system that includes a first function and a second function, or for analyzing a technical process that includes a first function and a second function, the device configured to:

• provide for the first function a first model, and provide for the second function, a second model is provided, the first model modeling a functional relationship of an input variable and an output variable of the first function and an influence of a first disturbance variable as a function of a first parameterization and as a function of a predefined learned distribution over the first parameterization, the second model modeling a functional relationship of an input variable and an output variable of the second function and an influence of a second disturbance variable as a function of a second parameterization and as a function of a predefined learned distribution over the second parameterization;
• provide an observation that includes input variables and output variables of the first function and of the second function;
• determine a prediction regarding an effect of:
• i) changing the first function under an influence of the first disturbance variable, and/or ii) changing the second function under an influence of the second disturbance variable;
• and
• select from the first function and the second function that function whose change has a higher likelihood of achieving a desired effect, compared to changing the other of the first function and the second function.

A non-transitory computer-readable medium on which is stored a computer program including computer-readable instructions for analyzing a technical system that includes a first function and a second function, or for analyzing a technical process that includes a first function and a second function, the instructions, when executed by a computer, causing the computer to perform the following steps:

• providing for the first function a first model, and providing for the second function, a second model is provided, the first model modeling a functional relationship of an input variable and an output variable of the first function and an influence of a first disturbance variable as a function of a first parameterization and as a function of a predefined learned distribution over the first parameterization, the second model modeling a functional relationship of an input variable and an output variable of the second function and an influence of a second disturbance variable as a function of a second parameterization and as a function of a predefined learned distribution over the second parameterization;
• providing an observation that includes input variables and output variables of the first function and of the second function;
• determining a prediction regarding an effect of:
• i) changing the first function under an influence of the first disturbance variable, and/or ii) changing the second function under an influence of the second disturbance variable;
• and
• selecting from the first function and the second function that function whose change has a higher likelihood of achieving a desired effect, compared to changing the other of the first function and the second function.