A Framework for Developing an Optimized Adaptive Scheduler for NASA Landolt Mission Observations

Scheduling observations for space missions is crucial for maximizing the quality and frequency of observations. Early scheduling efforts were entirely manual. Current scheduling techniques still require significant human intervention and have used various approaches, including reinforcement learning, neural networks, and metaheuristic algorithms. This paper presents a framework for optimizing an adaptive scheduler for the upcoming NASA Landolt mission which will send a satellite acting as an artificial star into orbit around the Earth. Observatories and telescopes are organized into a prioritization scheme based on features such as diameter, wavelength, and elevation. A long-term algorithm then uses the prioritization scheme to generate an initial ideal long-term schedule. As weather information and availability update, a short-term algorithm continuously adapts the schedule. Future simulations will be run on this framework to optimize the parameters of the algorithm.