qiskit-hackathon-korea-22
qiskit-hackathon-korea-22 copied to clipboard
qiskit-community
Q-urling (Quantum Curling)
Abstract
Define curling game in Markov Decision Process and find optimal strategy in the game using Qiskit Quantum Reinforcement learning.
Description
Curling is one of the favorite sports for spectators in the Winter Olympic game. Selecting a strategy is very important in playing curling (often called chess on ice). The entire curling game could be described into Markov Decision Process by expressing the strategy into aggressive and conservative ones[1]. Also, it is known that variational quantum circuits can perform reinforcement learning (policy-gradient) [2]. In this work, we will build a variational quantum circuit in Qiskit and train it to make this circuit decide the best strategy for playing curling.
Members
- @johnparkn
- @jaeunkim
- @chcy922
- @WestGround
- @anfry15rudals
- Qiskit Coach: @githubhandle
Reference
[1] Kiwook Beae, Dong Hyun Park, Dong Hyun Kim, and Hayong Shin, “Markov Decision Process for Curling Strategies,” Journal for Korean Institute of Industrial Engineers. Vol.42, No. 11, pp. 65-72, Feb 2016. [2] Sofiene Jerbi, Casper Gyurik, Simon C. Marshall, Hans J. Briegel, and Vedran Dunjko, “Parameterized Quantum Policies for Reinforcement Learning”, Advances in Neural Information Processing System 34, 2021.
