A quadratically convergent local algorithm on minimizing sums of the largest eigenvalues of a symmetric matrix (Q1312086)

The authors deal with the problem of minimizing sums of the largest eigenvalues of a symmetric matrix which depends on the decision variable affinely. Given \(\varepsilon\geq 0\) they derive an optimality condition which ensures that the objective function is within an \(\varepsilon\)- error bound of the solution. This condition may be used as a stopping criterion for any algorithm solving above problem. Also, one shows that in a neighbourhood of the minimizer the optimization problem can be equivalently formulated as a smooth constrained problem. For solving of the problem an algorithm which improves an other algorithm on minimizing the largest eigenvalue of a symmetric matrix is proposed. The convergence of the algorithm is illustrated by two numerical examples. One combines the algorithm with a simple technique and then global convergence is ensured.