Photonic decision making for solving competitive multi-armed bandit problem using semiconductor laser networks

Multi-armed bandit problems concern decision making when selecting a slot machine among many machines with initially uncertain hit probabilities to maximize the total reward; this is fundamental problem of reinforcement learning. Furthermore, competitive multi-armed involve multiple agents in play, manifesting concerns regarding social figures, not just individual rewards. A representative issue selection conflict, which players select same and may miss reward as whole. This study proposes scheme for solving using semiconductor laser networks by introducing an exclusive mechanism. We numerically implement our method compare it conventional algorithms. show that outperforms algorithms problem.