SlotMachine

An automatic conference schedule optimiser, built on the OR-Tools linear programming library and solver.

SlotMachine generates an optimal schedule for a multi-venue event, solving thousands of constraints on speaker availability and venue requirements in seconds. It can also take an existing schedule and calculate the minimum number of changes required to accommodate a change in constraints.

SlotMachine is used to generate the schedule for Electromagnetic Field events.

How to use

Create a SchedulingProblem out of a list of Talks:

from slotmachine import Talk, SchedulingProblem

talks = [
    Talk(
        # Talk ID for identifying talks in the result
        id=1,
        # Speaker IDs - the system will avoid scheduling two talks from the same speaker at the same time.
        speakers={1, 2},
        # Venue IDs
        allowed_venues={1},
        # Duration of the talk in minutes
        duration=30,
        # Time ranges when the talk is allowed to be scheduled
        allowed_times=[
            (datetime(2026, 5, 16, 12, 0, 0), datetime(2026, 5, 16, 19, 0, 0))
        ]
    ),
    ...
]

problem = SchedulingProblem(
    talks=talks,
    # The length of a "slot" in minutes - this is the granularity the scheduler will operate at.
    # All durations and allowed times must be multiples of this number.
    slot_duration=10
)

There are other options to the Talk class which are documented in the code. Now you can run the solver:

from slotmachine import SlotMachine

slotmachine = SlotMachine(problem)
result = slotmachine.solve()

result.talks will be a list of talks with the start_time and venue fields set.

Updating a schedule

If you need to update your schedule, you can pass a list of Talks with their existing start_time and venue values set. The solver will minimise the number of schedule changes required to accommodate the changes in constraints.

Venue/speaker availability

This library does not deal directly with venue or speaker availability, as this can be quite complex and event-specific. It can be handled outside SlotMachine by setting Talk.allowed_times to the intersection of the speaker and venue availability.

Per-venue time ranges are coming soon.

Acknowledgements

• The concept and code for the original version of this library was from David MacIver.
• For more information on this approach, see David's talk Easy solutions to hard problems from PyCon UK 2016.
• A similar library with a slightly different approach is conference-scheduler.