Mehrotra predictor–corrector method: Difference between revisions

Content deleted Content added
{{math-stub}} It will be nice if more detail is added. See gradient descent for an example.
{{math-stub}} It will be nice if more detail is added. See gradient descent for an example.
Line 3:
The idea is to first compute an optimizing search direction based on a first order term (predictor). The stepsize that can be taken in this direction is used to evaluate how much centrality correction is needed. Hence, a combined centrality term and second order term is computed (corrector).
 
[[Category:Algorithms]][[Category:Optimization]]
{{math-stub}}
[[Category:Algorithms]][[Category:Optimization]]