Satellite Observing Mission Scheduling Method Based On Case-based Learning

With the development of science and technology,major powers in the world is conducting deep research into the field of space exploration.Satellite as the main tool for space exploration,development prospects.Satellite plays the main role in the field of space exploration and has broad prospects for development.In the satellite earth observation,it is necessary to arrange the movement of the satellite for reasonable utilization of limited resources of satellites.Traditionally,in the process of the satellite earth observation,mathematical programming model is firstly established based on the characteristic of the problem,and then an intelligent optimization algorithm is selected for the solution.Traditional research is lack of reference to the historical scheduling cases.In the dissertation,case-based learning(CBL)method is introduced for the problem of satellite observing mission scheduling.The main research can be concluded as follows:1.The frame structure of satellite mission planning based on CBL is established.Due to the limitation in the model and algorithm of traditional satellite observing mission scheduling,CBL method is introduced and the frame structure of satellite mission planning based on CBL is established.2.Methods of case features extraction,case retrieval and case matching for satellite mission planning is proposed.The similarity between scenarios is analyzed in detail.And then the similarity criterion is designed as the basis of case feature extraction and case matching.It provides for the conduct of the similarity between scenarios from different levels,providing the strategies of CBL with the theoretical basis and operational means.3.The method of deconfliction in satellite missions and the evolutionary scheduling method both based on similar cases are proposed.A method of CBL is proposed to solve the problems of deconfliction in satellite missions.And the satellite observation mission planning method based on CBL-style genetic algorithm(GA)is proposed to solve the satellite observation scheduling problem with the advantages of CBL and GA.Finally,the feasibility and validity of the method are verified by simulation experiments.4.An experiment system for the practical application has been designed and implemented.The system validates the key points of the dissertation.