Q-adaptive routing improves system throughput and reduces packet latency in Dragonfly networks through multi-agent reinforcement learning.
