Multi-agent planning system for spacecraft

Recently, growing attention has been paid to multiagent planning system in artificial intelligence. A multiagent planning system (MAPS), which is used to produce applicable action sequence under complex constraints, is built for the spacecraft mission autonomous planning in this paper. The new architecture of distributed MAPS is designed, and the functions of the plan agents and manager agent are described. To be capable of describing simultaneous activity, continue time, resource and temporal constraints, a new planning formal model is given. For every planning agent in the MAPS, its layered architecture is designed and planning algorithm based on the temporal constraint satisfaction is given in detail. Finally, with the given scenarios, a prototype for the multi-agent planning system of spacecraft operation is implemented.