is a norm on the vector space M but is not a matrix norm. Consider the matrix J and compute J . It is not the case that , and hence is not a submultiplicative norm. However , if we define
then we have
Thus , only a minor modification of the vector norm || is required to make it a matrix norm.
Associated with each vector norm
\ on C is a natural matrix norm \ that is “induced” by \ on M . The norm \ is constructed from \ , and this construction adds to the list of methods for producing one norm from another.
The max in the above definition (rather than sup)
Is justified since || is a continuous function of x and the unit ball B is a compact set (see Appendix E).