The LCOGT network is scheduled as a single global entity. Requests are made for a specific telescope class (2m or 1m aperture), but do not identify specific sites or telescopes. The 1-meter telescopes are distributed, identical, and redundant. The LCOGT Global Adaptive Telescope Scheduler books each observation on the best telescope available, and rebooks it on the next available telescope in the event of failure due to weather or technical problems.
Requests are submitted, and the outcome of previous Requests inspected, via the LCO Observing Portal.
Each requests must have at least one target. The target's location is described with either RA/Dec coordinates or orbital elements (for moving targets). Each Request has one or more Time Windows. These are (potentially large) intervals of time during which the target is eligible for observation. For example, a Window of one week that begins immediately constrains the observation to be scheduled within the next week, or not at all.
Within the user-specified window, the scheduler automatically handles target observability (darkness and sky position), considering all times where the target can be observed anywhere across the network. Additional constraints, such as minimum airmass and distance from the moon may also be provided if necessary.
The scheduler is dynamic and highly responsive. When a new Request enters the system, the schedule is recomputed, and the new Request may be scheduled. The generation of a new schedule typically takes only a few minutes from submission, although the exact rerun time depends on multiple factors. Note that for efficiency reasons, observations currently taking place at a telescope are never interrupted before completion. Thus the earliest time from submission to observing at a telescope also depends on the time remaining for currently running observations. The only exception to this rule are Rapid Response (RR) observations, which disrupt running observations and will begin executing as soon as possible.