How AI can generate conflict-free school timetables in seconds

Building a school timetable is the kind of problem that looks simple until you actually try it. A 30-class CBSE school with 45 teachers, 8 periods a day, 6 working days, lab requirements, sports periods, substitute teachers, and a principal who’d like Friday afternoons free has astronomically many possible timetables. Most of them are illegal.

Done by hand, this exercise eats a senior teacher’s long weekend at the start of every term. EduPlux’s scheduler aims to do it in seconds — not because computers are clever, but because the problem is well-suited to the techniques we already have.

What it actually is

A timetable is a constraint satisfaction problem. Hard constraints must be satisfied (a teacher can’t be in two rooms at once). Soft constraints are preferences the optimiser tries to maximise (Mr. Iyer prefers afternoon slots; spread sports across the week). The job of the scheduler is to find a valid timetable that scores well against the soft preferences.

• Hard constraints: room conflicts, teacher conflicts, weekly subject hours, lab availability, fixed assemblies.
• Soft constraints: teacher preferences, balanced workload across days, no more than three consecutive periods of the same subject for a class.
• Custom rules: principals can pin specific periods, mark teachers unavailable on specific days, and lock entire blocks (e.g., Saturday closure).

Why this is now tractable

Modern constraint solvers (we use a combination of Google’s OR-Tools and a custom local-search refiner) make this dramatically faster than even ten years ago. A timetable that used to take a person a long weekend can now be produced — and re-produced when something changes — in well under a minute on commodity hardware.

The hard part isn’t the algorithm

The hard part, we’ve found, is data entry. A school has to express its rules clearly enough for a solver to use them. So a lot of our scheduler work is actually UX work: how do we let a school describe “Ms. Rao doesn’t teach two periods back-to-back” without learning a constraint language?

“Most of the value of an AI scheduler isn’t in the optimiser. It’s in the workflow that turns a principal’s instinct into something the optimiser can use.”

Where it’s headed

The first version, shipping with our launch, handles standard CBSE / ICSE / state board structures end to end. The next version will add multi-shift scheduling for institutions running morning and afternoon batches. If your school has an unusual scheduling rule, we’d love to hear about it — that’s exactly the kind of conversation that makes the next release better.