A Matrix Inversion Method Based on LDLT Decomposition and its Application in STAP

This paper presents matrix inversion algorithms based on LU decomposition and QR decomposition and LDLT decomposition (i.e. improved Cholesky decomposition) and the time complexity of the three methods is analysed. It is found that the inverse algorithm based on LDLT decomposition has the lowest complexity. Considering the resource consumption problem in practical engineering, this paper proposes the LDLT decomposition-based matrix inversion algorithm with partial double precision data type and the algorithm performance is simulated. It is finally found that the partial double precision algorithm for calculating the inverse matrix reduces resource consumption, improves operational efficiency and is within an acceptable margin of error from the true value, which is suitable for engineering Implementation.