Question: Numerical Methods Let A R mn , x R n and y = Ax. Assume x is representable in IEEE oating point arithmetic. Furthermore, let
Numerical Methods
Let A Rmn, x Rn and y = Ax. Assume x is representable in IEEE oating point arithmetic. Furthermore, let y^ (hat) represent the matrix vector product Ax computed in IEEE oating point arithmetic. Perform a backward error analysis to deduce an upper bound on the relative error.
||y^-y||2/||y||
Step by Step Solution
There are 3 Steps involved in it
Get step-by-step solutions from verified subject matter experts
