Method and device for learning a strategy and for implementing the strategy

authored by

Steven Adriaenssen, Andre Biedenkapp, Frank Hutter, Gresa Shala, Marius Lindauer, Noor Awad

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.

Organisation(s)

Machine Learning Section

External Organisation(s)

Robert Bosch GmbH

Type

Patent

Publication date

27.01.2022

Publication status

Published

Electronic version(s)

https://worldwide.espacenet.com/patent/search/family/079179186/publication/CN113971460A?q=CN113971460A (Access: Open)
https://worldwide.espacenet.com/patent/search/family/079179186/publication/US2022027743A1?q=US2022027743A1 (Access: Open)
https://worldwide.espacenet.com/patent/search/family/079179186/publication/DE102020209281A1?q=DE102020209281A1 (Access: Open)
https://worldwide.espacenet.com/patent/search/family/079179186/publication/JP2022022177A?q=JP2022022177A (Access: Open)
https://worldwide.espacenet.com/patent/search/family/079179186/publication/JP7695836B2?q=JP7695836B2 (Access: Open)