A Taxonomy of Descent Algorithms for Nonlinear Programs and 
Variational Inequalities

Michael Patriksson
Department of Mathematics
Linköping Institute of Technology
S-581 83 Linköping
Sweden

ABSTRACT:

We classify descent algorithms for nonlinear programs and variational
inequalities, based on their characterizations as cost approximation 
algorithms; an eight-tuple of parameters describes the problem and the choices
of cost approximating mappings and merit functions. The taxonomy is 
illustrated on classical algorithms and is utilized to interrelate known
algorithm frameworks.

KEY WORDS: Taxonomy, Classification, Nonlinear Programs,
Variational Inequalities, Descent Algorithms, Cost Approximation