Details
| Original language | English |
|---|---|
| Patent number | US2022027743 |
| IPC | G06N 3/ 12 A I |
| Priority date | 23 Jul 2020 |
| Publication status | Published - 27 Jan 2022 |
Abstract
A method for learning a strategy, which optimally adapts at least one parameter of an evolutionary algorithm. The method includes the following steps: initializing the strategy, which ascertains a parameterization of the parameter as a function of pieces of state information; learning the strategy with the aid of reinforcement learning, it being learned from interactions of the CMA-ES algorithm with a parameterization, determined with the aid of the strategy as a function of the pieces of state information, with the problem instance and with a reward signal, which parameterization is optimal for possible pieces of state information.
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
Patent No.: US2022027743. Jan 27, 2022.
Research output: Patent
}
TY - PAT
T1 - METHOD AND DEVICE FOR LEARNING A STRATEGY AND FOR IMPLEMENTING THE STRATEGY
AU - Adriaenssen, Steven
AU - Biedenkapp, Andre
AU - Hutter, Frank
AU - Shala, Gresa
AU - Lindauer, Marius
AU - Awad, Noor
PY - 2022/1/27
Y1 - 2022/1/27
N2 - A method for learning a strategy, which optimally adapts at least one parameter of an evolutionary algorithm. The method includes the following steps: initializing the strategy, which ascertains a parameterization of the parameter as a function of pieces of state information; learning the strategy with the aid of reinforcement learning, it being learned from interactions of the CMA-ES algorithm with a parameterization, determined with the aid of the strategy as a function of the pieces of state information, with the problem instance and with a reward signal, which parameterization is optimal for possible pieces of state information.
AB - A method for learning a strategy, which optimally adapts at least one parameter of an evolutionary algorithm. The method includes the following steps: initializing the strategy, which ascertains a parameterization of the parameter as a function of pieces of state information; learning the strategy with the aid of reinforcement learning, it being learned from interactions of the CMA-ES algorithm with a parameterization, determined with the aid of the strategy as a function of the pieces of state information, with the problem instance and with a reward signal, which parameterization is optimal for possible pieces of state information.
M3 - Patent
M1 - US2022027743
ER -