How did you model the VRP as a QUBO? Share your formulation and trade-offs

We're interested in how participants approached the optimization problem. Some discussion prompts: How did you formulate the Vehicle Routing Problem (VRP) as a QUBO? Which penalty terms worked best for capacity, time windows, and route constraints? Did you prioritize solution quality, quantum circuit efficiency, or execution time? Which classical heuristics (e.g., greedy, 2-opt, local search) did you combine with your quantum approach? What trade-offs or unexpected challenges did you encounter? Feel free to share your ideas, design decisions, or lessons learned without revealing code or proprietary implementations. Let's compare different hybrid quantum-classical strategies and learn from each other's approaches.