The average complexity of the case is quite difficult to analyze, and it depends on the distribution of your linear program. I believe that it was designed as polynomial time within some common distributions. Currently I can not find the article.
EDIT : Yes, here are the sources:
Nocedal, J. and Wright, SJ Numerical Optimization. New York: Springer-Verlag, 1999.
Forsgren, A .; Gill, PE; and Wright, M. H. "Internal Methods of Nonlinear Optimization." SIAM Rev. 44, 525-597, 2002.
I read it in the first book and, apparently, it was proved in a separate article (Forsgren). You can find either in the university library.
tskuzzy
source share