I already have a working-ish vehicle routing model written in Python that relies on heuristic solvers, but the algorithm execution layer keeps throwing unexpected results and occasional crashes. The data ingestion and output formatting steps behave, so I’m certain the fault sits squarely in the core optimisation loop. Here’s what I need from you: • Step through the current codebase with me, identify the exact logic or state-handling errors that cause the routes to explode or the solver to hang. • Refactor or patch the offending blocks so the solver converges consistently and produces viable routes for the daily test dataset. • Add concise inline comments and, if a third-party optimisation library (e.g., OR-Tools, Pyomo, NetworkX, or your preferred heuristic framework) can simplify or stabilise the workflow, wire it in. Acceptance criteria – Same input now returns a complete, non-crashing set of routes within expected runtime. – Unit tests (I’ll supply sample cases) all pass. – Code remains Pythonic, readable, and easy for me to extend. I’ll grant repo access as soon as we agree. If you’ve wrestled with VRP, TSP variations, or similar meta-heuristic implementations before, this should be straightforward.