Automated Planner for a Field-Service Maintenance Schedule
Overview
What this challenge is about.
Model the scheduling problem as either PDDL (Planning Domain Definition Language) or a constraint-satisfaction problem using a solver like OR-Tools. Define: actions (visit_site, swap_technician, defer_visit), preconditions (skill match, parts available, within visit window, not in customer blackout), and effects. For PDDL: use Fast Downward as the planner. For CSP: use OR-Tools CP-SAT. Run on 1 month of anonymized real data. Evaluate against the human planner on: % visits scheduled, drive-time per visit, customer-blackout violations, deferred-visit count. Deliver: planner repo + tests, evaluation report (6 pages), 1-page ship recommendation.
The Brief
What you'll do, and what you'll demonstrate.
Build a classical-AI planner for preventive-maintenance scheduling that matches or beats the human planner on visit-coverage + drive-time efficiency.
Earning criteria — what you'll demonstrate
- Apply classical AI planning (PDDL or CSP) to a real scheduling problem
- Design an encoding that captures domain constraints faithfully
- Evaluate AI planners against human-decision baselines defensibly
- Recommend ship/don't-ship with trade-off honesty
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.
Product Manager
PMs on field-ops and logistics products need this planner fluency to scope automation that respects real-world constraint complexity.
This challenge sharpens
- planning
- algorithm-evaluation
- intelligent-agents