Community Cadence: Adaptive Queue Scheduling for the Keck Planet Finder
Abstract
Precise Doppler studies of extrasolar planets require fine-grained control of observational cadence, i.e. the timing of and spacing between observations. This enables optimal sampling of planetary and stellar activity signals, spanning hours to decades. Achieving acceptable cadence at classical observatories has been a challenge for the past three decades. As next-generation Doppler instruments achieve systematic errors well below 1 m/s, control of observational cadence will become even more critical given its central role in stellar activity mitigation techniques. One such instrument is the Keck Planet Finder (KPF) with an engineered instrumental stability of ~30 cm/s that is being commissioned at W. M. Keck Observatory (WMKO) on Maunakea. We present a novel scheduling framework called 'Community Cadence' which balances the cadence needs of KPF with WMKO's existing classical operations model. For a set of observing programs and allocated nights, our software seeks to determine the optimal timing and ordering of ~1000 observations within a given observing semester. We achieve a near-optimal solution in real-time using a hierarchical Integer Linear Programming (ILP) framework. ILP solvers are used throughout the business operations and logistics sciences to solve high-dimensional optimization subject to large numbers of constraints. Community cadence optimizes over the ~103000 possible orderings. A top level optimization finds the most uniform set of allocated nights by which to observe each host star in the request catalog. A second optimization schedules the observations assigned to each night in the manner that minimizes telescope slews. The resulting schedule not only maximizes program completion, but concurrently improves planet detection and activity signal mitigation.
- Publication:
-
American Astronomical Society Meeting Abstracts
- Pub Date:
- January 2023
- Bibcode:
- 2023AAS...24126702H