Research On Comprehensive Evaluation Of Satellite Mission Scheduling Algorithms

Satellites obtain the earth's image data from space through on-board sensors, which has become an important means of investigating and researching the earth's resources. In recent years, in order to make better use of the existing satellite resources, various algorithms have been proposed and applied in satellite mission scheduling, which achieved some fruitful results. When scheduling, in face of all these algorithms, how to comprehensively evaluate and reasonably choose them has becoming a pressing problem. In addition, as the aerospace industry develops, the practical applications change constantly, which sets a higher demand to the further extension of the scheduling algorithms.At present, some researchers have done a lot of work in comparing and evaluating the performance of the existing satellite scheduling algorithms, which to some extent is meaningful. But, due to the lack of standard test data, unified measurements and scientific evaluation methods, the existing comparison and evaluation could not directly guide selecting algorithms while scheduling and extending them forward. As a result, to deeply study the comprehensive evaluation technology of satellite mission scheduling algorithms is the objective demand of scientific development on satellite mission scheduling, which has great theoretical and practical significance.To this end, this thesis studied the key technologies in example automatic generation, measurements architecture, comprehensive evaluation, and built a comprehensive evaluation scheme of satellite mission scheduling algorithms. The main work and innovations are as follows:1. Based on deep analysis of the satellite mission scheduling problem, the limitations of existing comparison and evaluation were pointed out, and then, a comprehensive evaluation scheme was put forward.2. After researched a great deal of real satellite observing targets, the'Satellite Observing Target Model'was built under the mature statistics theory. Based on this model, satellite observing targets, i.e. the test examples, could be automatic generated according to different evaluation purposes.3. With various aspects of algorithm's performance abstracted, the corresponding evaluation measurements were formed. Then, according to the AHP algorithm, these evaluation measurements were divided into different levels, and a high-scalable evaluation measurements architecture was built, with its rationality, redundancy, and integrity thoroughly discussed.4. Considering both subjective and objective experience and knowledge, the scientific comprehensive methods were put forward, through which the single measurements were synthesized, and the evaluation conclusion was formed. According to the conclusions above, algorithm automatic selection, based on Rough Set Theory, was implemented, and the possible extending direction was pointed out.5. Based on the comprehensive evaluation scheme, a mission scheduling algorithm evaluation-management system was implemented. The evaluation conclusions gained by this system could be directly employed in selecting algorithms while scheduling and extending algorithms in the future.