Planner · simulation

More robots do not make a faster night.

A discrete-event model of one cleaning night in a tall building. Change the building, the fleet and the elevator; the planner finds the robot count that finishes first. Every input is an assumption until a pilot measures it.

Model v0.1runs in your browsernothing is sent or stored2026-10-10

01 — Plan a night

Defaults: the worked example on the home page
Building
Fleet
Elevator
Plan
90robots
45400

With 90 robots, staggered starts end the night at 01:00.

Fits the window
Night, staggered starts
3.0h
done 01:00
Simulated
Night, all up then clean
4.0h
done 02:00
Simulated
Elevator trips each way
30trips
per elevator
Simulated
Robots left free
310robots
for other jobs that night
Simulated
Fig. 2 The chosen plan against sending every robot, from 22:00
Chosen count, staggered starts Chosen count, all up then clean Every available robot, all up then clean Elevator busy Robots cleaning Window ends

Cyberstrata planner v0.1discrete-event modelinputs as set above

Fig. 3 Length of the night against robots deployed, h
Staggered starts All up, then clean Window ends

Click the chart or use ← → to choose a robot countSimulated

02 — The model

Planner v0.1 · rules and limits

The elevator is a shared queue. The plan decides who rides it, and when.

Two schedules run on the same inputs. All up, then clean is how the worked example is usually told. Staggered starts let each robot clean as soon as it arrives and queue for the ride down when its area is done.

  1. 01

    Robots are split evenly across floors, at least one per floor, and stay on their floor all night.

  2. 02

    A floor's area is split evenly among its robots. Robots with the most area ride up first.

  3. 03

    Each elevator trip takes one round trip and carries up to the set number of robots. Trips up are served before trips down.

  4. 04

    The night ends when the last robot is back at the loading dock.

Not modelled yetEffect on the night
First-time mapping of each floorLonger first night; repeats unaffected
Charging and battery limitsCaps area per robot per night
Transport to the siteEarlier start, favours local fleets
Elevators shared with peopleLonger and less regular trips
Toilets, desks and binsDone by people in a mixed crew
Faults and stopsNeeds spare robots in the plan

Next: a matching simulation in which jobs with real constraints meet robots with real limits, the fit check filters what cannot work, and the matches feed a track record.