Multiple computations of capacitance matrix of conductor and dielectric system with varying parameters by block Lu-decomposition

This work considers a set of problems concerned on multiple computations of systems of linear algebraic equations by using block LU-decomposition. We found out matrix specificity, which can be used for mentioned purpose, and analyzed a role or new iteration renumbering that provides much to improve solution efficiency and to increase range of analyzed structures under low computing cost. We improved algorithm for computing set of capacitance matrixes of conductor and dielectric system and derived analytical expressions for evaluation of arithmetic cost. These expressions provide to perform a-priory estimation of block LU-decomposition using. We produced analytical and computational estimations of speedup of the algorithm depending on different parameters. The maximal estimated value of speedup is 23 times.