Twenty Dinners, One Hill

What you will do

Design and improve the algorithms, one constraint at a time. The problem grows in five steps: find a route (weeks 1–3); find a route the drone can actually fly and return from (week 4); order many deliveries for one drone (weeks 5–6); divide the work between drones that share two charging pads (weeks 7–8); make the plans of five drones executable together in one shared airspace (weeks 9–12). Each step starts from the method you already have, breaks it on a small example, and makes you change the state, the cost or the check that was wrong.

Every week you write a prediction before you run anything, derive one thing by hand, implement one key function, test it against a counterexample, and explain the result under the objective the course fixed. You leave with a cumulative planner — search, scheduling, space-time coordination — and the arguments, tests and records that say what it guarantees and what it does not.

• planning
• drones
• algorithms

Who it is for

Third-year undergraduates who have done introductory programming and data structures: functions, loops, lists and dictionaries, queues and priority queues, basic graphs, and enough algebra to sum costs along a path. No robotics, no flight control, no prior optimisation course. If you can write a loop that pops the smallest item from a queue, you can start; if you want a course about real drones and real airspace rules, this is not it — every number here is a teaching setting, and the policies page says so.