Build Real-time Operational Loop and Monitoring

Implement a continuous operational loop where the Mastra AI agent periodically receives new requests and fleet updates, generates plans, validates them, and updates the simulated fleet status. Focus on ensuring low-latency responses, leveraging the TensorRT-LLM optimization. Implement basic logging to monitor agent decisions and model response times.