Adaptive wireless resource selection is essential in dynamic radio environments with rapidly changing traffic conditions. This study applies ultrafast decision-making based on laser chaos to wireless resource selection, implementing and comparing two frameworks: the Laser Chaos Decision Maker and the Laser Network. The former leverages the chaotic dynamics of a single semiconductor laser, while the latter utilizes optically coupled lasers and chaos synchronization for scalable selection. Simulations using a virtual network environment evaluate their performance in IEEE 802.11ax channel selection under static and dynamic conditions. Results show that both methods outperform conventional multi-armed bandit algorithms in terms of correct decision rate (CDR). Additionally, incorporating memory effects into Q-value updates enables adaptation to time-varying environments. These findings highlight the potential of laser chaos-based decision-making for next-generation wireless systems.
