Computes eigenvalues and eigenvectors of a given symmetric tridiagonal matrix T using
"Algorithm of Multiple Relatively Robust Representations" (MRRR).
"Dhillon, Inderjit S. and Parlett, Beresford N., "Multiple representations to compute
orthogonal eigenvectors of symmetric tridiagonal matrices", Linear Algebra and its Applications,
2004, 387, pp. 1-28."
"Marques, Osni A., Riedy E. Jason, and Vomel Christof, "LAPACK working note 172: Benefits of
IEEE-754 features in modern symmetric tridiagonal eigensolvers", Technical Report UCB//
"Dhillon, Inderjit S., Parlett, Beresford N. and Vomel, Christof, "The design and
implementation of the MRRR algorithm", ACM Transactions on mathematical Software,
December 2006, Vol. 32 No.4, pp.533-560."
public static final double DEFAULT_MIN_RELATIVE_GAP
Default value for the minimum relative gap threshold. This threshold is used to determine if
two eigenvalues are too close.
When a child is inside a tight cluster, it can be difficult to find an RRR. A partial remedy
from the user's point of view is to make this threshold smaller and recompute.
However, as the orthogonality of the computed vectors is proportional to the reciprocal of
this threshold, decreasing the value of this threshold will also decreases the precision of
the computed eigenvectors.