A generalization of the numerical range of a matrix (Q923679)

The classical numerical range and some known generalizations are stated first, then a new generalization of the classical range of a matrix is introduced. The main theorem is that four conditions on the numerical range are equivalent to each other. Various properties and basic characterizations of this generalized numerical range are established. One of these characterizations is used in proving a resolvent inequality for a discretized version of the differential operator \(d^ 2/dx^ 2\).