Artificial intelligence scheduling for NASA's Hubble Space Telescope
Mark Johnston, Glenn E. Miller · 2002
An artificial intelligence (AI) system called Spike has been implemented for scheduling observations with NASA's Hubble Space Telescope. The system incorporates innovative methodologies for representing and reasoning with scheduling constraints and preferences and for conducting scheduling search. For the former, a combination of constraint satisfaction techniques and weight-of evidence combination has been devised to propagate temporal constraints and preferences. For the latter, a neural network scheduling has been found to be highly efficient. The Spike system is operational and is being used to generate long-range schedules for Space Telescope operations to follow initial on-orbit checkout.>