CRPC-TR92022:  Solving Nonlinear Integer Programs with a Subgradient 
Approach on Parallel Computers

Robert Bixby, John Dennis, Zhijun Wu

June, 1992

Many large, and hard, nonlinear integer programming problems-or more 
generally mixed-integer nonlinear programming problems-arise in both 
theoretical study and practical applications. No general and efficient 
solution to these problems can be found with traditional computers. Thus, 
it becomes necessary to develop new algorithms for solving these problems 
on advanced architectures, such as parallel computers. This article 
describes a new algorithm, developed at the Center for Research on 
Parallel Computation at Rice University, for solving nonlinear integer 
programs and shows how the solution to a nonlinear integer program can be 
achieved on a parallel computer.