Research on constellation fault satellite replacement strategy based on Markov decision process

With the decrease of the importance of single satellite caused by the large-scale constellation，multiple replacement methods and the retention of faulty satellites should be considered when formulating the replacement strategy of faulty satellites. In response to this requirement，a method of formulating constellation orbital satellite replacement strategy based on Markov decision process is proposed. The orbital states are divided to the number of satellites in orbit，on-orbit backup and ground backup. On this basis，a return function model representing the expected maintenance cost is constructed by combining different replacement methods of faulty satellites and corresponding ordering，storage and penalty costs. Sequential replacement strategies are formulated with the goal of minimizing the expected maintenance cost of each decision-making cycle and an application analysis is carried out taking a certain constellation orbit as an example. The results show that when ground backup satellites are used to replace faulty satellites，it tends to supplement the consumed. The longer the mean time between failures，the lower the total expected maintenance cost，but the effect is gradually weakened. The increase of penalty cost will lead to the solidification of strategies，and the upward trend of total maintenance cost is expected to be greatly slowed down.