Method, controller, and computer product for reducing oscillations in a technical system

A controller for reducing oscillations in a technical system, the controller being configured to perform a method for configuring the technical system and reducing oscillations, the method comprising:

• a) receiving a multitude of different control policies for the technical system, wherein each respective control policy results from training one or more control models for the technical system by using one or more machine learning methods, wherein the multitude of different control policies includes at least one hundred different control policies,
• b) controlling the technical system with each of the respective multitude of different control policies,
• c) receiving a respective signals, one for each of the multitude of different control policies, each of the respective signals representing a time series of operational data of the technical system controlled by a respective one of the multitude of different control policies,
• d) processing each of the respective signals, the processing comprising a transformation into a frequency domain, wherein the processing further comprises determining an autocovariance function, and
• e) determining an entropy value for each of the multitude of different processed signals,
• f) comparing the determined entropy value for respective control policies, wherein comparing includes comparing the determined entropy value for respective control policies with determined entropy values for other respective control policies, and selecting a control policy controller setting from the multitude of different control policies depending on the comparing, and
• g) outputting the selected control policy controller setting and configuring the technical system by permanently implementing the outputted selected control policy controller setting in a controller of the technical system.

A computer program product, comprising a computer readable hardware storage device having computer readable program code stored therein, said program code executable by a processor of a computer system to implement a method for configuring a technical system and reducing oscillations the method comprising:

• a) receiving a multitude of different control policies for the technical system, wherein each respective control policy results from training one or more control models for the technical system by using one or more machine learning methods, wherein the multitude of different control policies includes at least one hundred different control policies,
• b) controlling the technical system with each of the respective multitude of different control policies,
• c) receiving a respective signals, one for each of the multitude of different control policies, each of the respective signals representing a time series of operational data of the technical system controlled by a respective one of the multitude of different control policies,
• d) processing each of the respective signals, the processing comprising a transformation into a frequency domain, wherein the processing further comprises determining an autocovariance function, and
• e) determining an entropy value for each of the multitude of different processed signals,
• f) comparing the determined entropy value for respective control policies, wherein comparing includes comparing the determined entropy value for respective control policies with determined entropy values for other respective control policies, and selecting a control policy controller setting from the multitude of different control policies depending on the comparing, and
• g) outputting the selected control policy controller setting and configuring the technical system by permanently implementing the outputted selected control policy controller setting in a controller of the technical system.