Simulator Design
The LSST Operations Simulator is based on an open-source simulation package--SimPy. SimPy is an object-oriented , process-based discrete-event simulation package based on standard Python and released under the GNU GPL.
The simulator is modular in design and will accept multiple, distinct observing proposals. Each proposal ranks potential observations based upon user-specified parameters, but hard-coded algorithms. Rankings are based upon current seeing, sky brightness, and progress towards completing a proposal based on previous observations. For proposals with a specified cadence, the proposal increases rankings for observations useful for that cadence, but does not look ahead to determine future effects such as fields setting.
Modeling the Telescope and the Sky
The simulator uses a sophisticated sky model, calculating sky brightness using the Krisciunas and Schaeffer (1991) model. It tracks the sun and moon using SLALIB routines.
A detailed telescope model is used to calculate slew time penalties. These penalties factor into the scheduling decision. All movements are tracked: mount, dome, optics, instrument rotator, cable wraps and filter changer. The velocities and accelerations for these motions are all settable parameters. There are open-loop optics alignments for all moves and closed-loop alignments for moves in altitude greater than a settable parameter (currently 9 degrees).
The simulator models seeing and cloud data. Seeing data has been generated using available DIMM data to determine the power spectrum of the seeing throughout the year and creating complete data sets with that power spectrum. Cloud data is from 10 years of CTIO night assistant observations. The simulator currently assumes alternating one and two week shutdowns per year for maintenance.
Analysis and Reporting
More than 200 user-defined parameters govern the telescope model, science program constraints, scheduling controls, and site characteristics. These parameters permit investigation of the effect of changes on the survey, in effect tuning the survey. More work is needed to explore the parameter space.
Simulation results are analyzed through the use of post-processing tools that culminate in a ~60-page report called the Simulated Survey Technical Analysis Report (SSTAR).
Next Steps
Work is progressing on a new code version of the simulator. Reorganizing the code will permit experimentation with different algorithms for scheduling observations. The new version schedules ahead by a user-defined amount of time, permitting schedule optimization over that time. In addition, it can schedule from a list of RA, Dec, filter, and exposure time generated by another program or a previous simulation. Assessing the effect of small changes in parameters on the simulated survey will be simplified by using previous outputs as input.
Source: American Astronomical Society 213th Meeting, Poster Exhibit, "LSST: Cadence Design and Simulation", K.H. Cook et al., 460.04