发明申请
WO2022013125A1 VERFAHREN ZUM KONFIGURIEREN EINES STEUERUNGSAGENTEN FÜR EIN TECHNISCHES SYSTEM SOWIE STEUEREINRICHTUNG
审中-公开
基本信息:
- 专利标题： VERFAHREN ZUM KONFIGURIEREN EINES STEUERUNGSAGENTEN FÜR EIN TECHNISCHES SYSTEM SOWIE STEUEREINRICHTUNG
- 专利标题（英）：METHOD FOR CONFIGURING A CONTROL AGENT FOR A TECHNICAL SYSTEM, AND CONTROL DEVICE
- 申请号：PCT/EP2021/069269 申请日：2021-07-12
- 公开(公告)号：WO2022013125A1 公开(公告)日：2022-01-20
- 发明人： RUNKLER, Thomas , SWAZINNA, Phillip , UDLUFT, Steffen
- 申请人： SIEMENS AKTIENGESELLSCHAFT
- 申请人地址： Werner-von-Siemens-Straße 1
- 专利权人： SIEMENS AKTIENGESELLSCHAFT
- 当前专利权人： SIEMENS AKTIENGESELLSCHAFT
- 当前专利权人地址： Werner-von-Siemens-Straße 1
- 优先权： EP20185973.3 2020-07-15
- 主分类号： G06N3/00
- IPC分类号： G06N3/00 ; G06N3/04 ; G06N3/08 ; G06N5/00 ; G06N7/00
To configure a control agent (POL), predefined training data are read in, which specify state datasets (S), action datasets (A) and resulting performance values (R) of the technical system (TS). Using the training data, a data-based dynamic model (NN) is trained to reproduce a resulting performance value (R) using a state dataset (S) and an action dataset (A). An action evaluation process (VAE) is also trained to reproduce the action dataset (A) using a state dataset (S) and an action dataset (A) after an information reduction has been carried out, wherein a reproduction error (DR, DO, Dl) is determined. To train the control agent (POL), training data are supplied to the trained dynamic model (NN), the trained action evaluation process (VAE) and the control agent (POL). Performance values (R1, R2) output by the trained dynamic model (NN) are fed into a predefined performance function (P). Reproduction errors (DO, Dl) output by the trained action evaluation process (VAE) are also fed as performance-reducing influencing variables into the performance function (P). The control agent (POL) is thus trained to output an action dataset (A) optimising the performance function (P) on the basis of a state dataset (S).
IPC结构图谱:
| G | 物理 |
| --G06 | 计算；推算；计数 |
| ----G06N | 基于特定计算模型的计算机系统 |
| ------G06N3/00 | 基于生物学模型的计算机系统 |