Design a room allocation service that, given a meeting request (time window, duration, capacity, equipment, location), returns a ranked list of rooms. Propose a scoring function integrating factors such as historical room usage counts, expected idle-time fragmentation, attendee walking distance, equipment match, past no-show rates, and fairness across teams. Describe data sources and freshness, how weights are tuned (rules vs learned model), safeguards against starvation, online/offline evaluation, and an A/B testing plan.