An effective Genetic Programming Hyper-Heuristic for Uncertain Agile Satellite Scheduling
A new AI system uses genetic programming to create real-time satellite task plans that outperform human-designed methods.
Researchers Yuning Chen, Junhua Xue, and team developed a Genetic Programming Hyper-Heuristic (GPHH) for the Uncertain Agile Earth Observation Satellite Scheduling Problem (UAEOSSP). The AI automatically generates scheduling policies that adjust plans in real-time for uncertain factors like task profit and visibility. In tests, these evolved policies achieved an average 5.03% improvement over Look-Ahead Heuristics and 8.14% over Manually Designed Heuristics, enabling more efficient satellite operations.
Why It Matters
This means satellites can be scheduled more efficiently and adaptively, maximizing valuable observation time and data collection for Earth monitoring.