Next: Error Calculation from
Up: 4: Minimization
Previous: 4.3The Derivative Method
- Now given any guess
, expand
.

- Adopt a vector notation with
.

- Minimum condition is

- So we have new guess

for true minimum.
- Note that as M positive definite in Equation (30),
this new value of
indeed gives a
minimum (not a maximum or saddle point), as long as second order
Taylor expansion valid.
Geoffrey Fox, Northeast Parallel Architectures Center at Syracuse University, gcf@npac.syr.edu