Overview
What this challenge is about.
You receive a week of anonymized daily VRPTW instances (around 800 orders per day, 120 vehicles, hard delivery windows). Implement tabu search with: a route-insertion construction heuristic for the initial solution, swap and relocate neighborhoods, aspiration criteria, and an adaptive tabu-tenure schedule. Benchmark against Google OR-Tools' VRP solver on solution quality (total distance + lateness penalty) and wall-clock time at a 60-second budget per instance. Report per-instance results and write a memo on whether the custom solver is worth maintaining.
The Brief
What you'll do, and what you'll demonstrate.
Build a tabu-search VRPTW solver and quantify whether it beats Google OR-Tools on the operations team's real instance distribution at a 60-second budget.
Earning criteria — what you'll demonstrate
- Implement tabu search with neighborhoods, aspiration, and tenure scheduling
- Apply metaheuristics to a constrained real-world routing problem
- Benchmark a custom solver against an industrial baseline
- Reason about the maintain-vs-replace trade-off for OR tooling
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
Applied AI Scientist
Building a custom metaheuristic for a real operations problem and quantifying it against a commercial baseline is core applied-AI-scientist work at any logistics or delivery firm.
This challenge sharpens
- tabu-search
- vehicle-routing
- optimization
Data Scientist
Per-instance gap analysis and a maintain-vs-replace memo are exactly the deliverables a data scientist ships to an operations leadership team.
This challenge sharpens
- benchmarking
- optimization
- python
Machine Learning Engineer
Writing tested, profiled Python that respects a runtime budget is the MLE craft of shipping algorithms that survive production.
This challenge sharpens
- python
- benchmarking
- metaheuristics