Wednesday, November 11, 2009

Persistency in math programming models

For a model I am working on we need a user settable degree of persistency: solutions should take into account historic solutions and stay “close” to those in some sense. Basically: wildly different solutions are unwanted. An excellent paper on this subject is by the NPS people: http://faculty.nps.edu/dell/docs/optimization_and_persistence.pdf.