Abstract: In response to the characteristics of a large number of remote satellite nodes, large task demands, complex resource utilization constraints, and high requirements for modeling and solving collaborative scheduling between multiple nodes in large-scale constellation coordination scheduling, the paper proposes a top-level framework for phased unified modeling and solving, which includes “task pre-processing; unified modeling; optimization solving; on-orbit instruction generation”. Under the top-level framework, a multi node real-time collaborative scheduling algorithm based on improved contract network (CSA-ICNP) is proposed for space-based early warning application scenarios, which utilizes fuzzy optimization combined with local search strategy to improve the overall optimization ability of the algorithm. The correctness and effectiveness of the proposed algorithm were verified by conducting a large number of simulation experiments and comparing the results with RS, GS, TAACD, MLFA and DSRS-ICNP that each experimental case achieves the optimal objective function value performance with an average improvement was 42.13%, 41.51%, 37.93%, 37.53% and 18.57%.

Key words: Large-scale constellation, Collaborative scheduling, Normalized modeling, Space-based warning, Improved contract net protocol