Poster Abstract

P8.11 Marco Iacobelli (ASTRON)

Theme: Telescope operations and scheduling: from classical to autonomous

The LAST resort: towards automatic scheduling of the LOFAR telescope.

Scheduling the observing program of a facility is one of the most important tasks in science operations of an observatory. Moreover, the new observing capabilities offered by current and next generation massive arrays pose a challenge in translating user proposed campaigns into an efficient observing program. A robust automatic (and dynamic) scheduling is therefore fundamental to increase the instrument observing time and its scientific output.

However, the increased complexity of such instruments make the scheduling exercise even more complex. In particular, this is true for distributed telescopes (i.e. LOFAR, SKA). For example, in addition to observing hours scheduling has to take into account the processing and archiving resources needed as they are generally limited given the large data rates and volume involved. However, as some of these constraints can be evaluated only when the observing program is executed, dynamic scheduling is essential.

In this contribution we present our analysis of the problem in the framework of LOFAR operations and we present a preliminary tool (LAST) to forecast the user proposed observing campaigns in an efficient observing program, thus satisfying various constraints for the proposed observations.

The tool, available on github, can already be used to advice the LOFAR time allocation committee, interactively. The tool is based on the astroplan library and it can calculate and plot for each observing block the possible observing date. Currently, the LOFAR scheduling task is still performed manually, but the next step is to include automatic scheduling in LAST. Finally, we aim for dynamic scheduling, where the next observation is picked based on changing constraints (e.g. ionospheric conditions, triggered ToO, etc.).