Multi-satellite multi-point target observation task planning method based on improved genetic and firefly combinatorial algorithm

Abstract

The invention relates to a multi-satellite multi-point target observation task planning method based on an improved genetic and firefly combinatorial algorithm. According to the method, an existing genetic algorithm is improved, the improved genetic algorithm can jump out of a non-feasible solution area more quickly, a feasible solution is found, and then a subsequent selection process is entered; then the improved genetic algorithm is combined with a firefly algorithm, the firefly algorithm has good locality in the later period, the optimization capacity of the genetic algorithm can be obviously improved, the search capacity of the algorithm near the optimal solution is improved, meanwhile, more cross processes occur in better individuals, and the algorithm performance is obviously improved; in addition, a non-machine learning analogue simulation optimization scheme is adopted, the large demand of satellite planning for previous planning data is avoided, it is guaranteed that the planning method is not affected by previous poor planning schemes, the constraint of time can be got rid of through the simulation technology, decision making and evaluation are conducted on satellite planning in advance, and cost is reduced.

Description

technical field [0001] The invention belongs to the field of satellite observation, in particular to a multi-satellite multi-point target observation task planning method based on an improved genetic and firefly combination algorithm. Background technique [0002] In the study of satellite planning problems, it can be roughly divided into two categories according to the methods used: research methods using intelligent optimization algorithms and research methods using deep learning for optimization. If it is required to make full use of satellite observation resources and allocate observation resources reasonably according to priorities, it is necessary to use intelligent optimization algorithms to solve this type of problem. Because the solution of deep learning is to use the historical scheduling data of a given satellite, train a prediction model based on augmented topology collaborative neuroevolution to predict the probability of different satellites completing the miss...

AI Technical Summary

Problems solved by technology

However, in practical applications, due to the characteristics of multi-satellite multi-point target observation mission planning problems, that is, most solutions in the solution space are infeasible solutions, so a simple algorithm often cannot obtain better results, and a large number of invalid results will be obtained. solution, even if an effective solution is obtained, the efficiency of a certain process of the algorithm itself is not high, and the availability of results will be reduced if the optimization function is not fully utilized

[0004] Genetic algorithm is the most feasible method for satellite observation task planning, but when it is applied to this problem, its own non-directional variation flow cannot play a good role in optimization, and as the algorithm gradually converges, the algorithm Poor search performance in local area

Images