2022/03/01 by Lloyd Nesbitt, Nesbitt, Lloyd
Computer Science · Engineering · Mathematics · #Matrix Theory and Algorithms #graph theory and CDMA systems #Advanced Optimization Algorithms Research
paper · pdf · doi:10.48550/arxiv.2203.00177
Given an n × n nonsingular matrix A and the characteristic polynomial of A as the starting point, we will leverage the Cayley-Hamilton Theorem to efficiently calculate the maximal length Jordan Chains for each distinct eigenvalue of the matrix. Efficiency and speed are gained by seeking a certain type of starting vector as the first step of the algorithm. The method for finding this starting vector does not require calculating the ker[(A-λI)k] which is quite an expensive operation, and which is the usual approach taken in solving the Jordan Canonical basis problem. Given this starting vector, all remaining vectors in the Jordan Chain are calculated very quickly in a loop. The vectors comprising the Jordan Chains will then be used to minimize the amount of equation solving in order to find the remaining generalized eigenvector basis. We will prove a theorem that justifies why the resulting Jordan Chains are of maximal length. We will also justify how to derive the starting vector which will subsequently be used to calculate the Maximal Jordan Chains.