The Solution of the Chemical Equilibrium Programming Problem with Generalized Benders Decomposition

This paper deals with an algorithm for the solution of the chemical equilibrium problem that parallels Geoffrion's development of the Generalized Benders Decomposition Algorithm. The unknowns of the problem are decomposed into the phase sums, which are the complicating variables, and normalized variables. The algorithm obtained is an iterative procedure where the master problem computes optimal phase sums and the subproblem computes new dual variables. A numerical study indicates that best results are obtained when the subproblem is suboptimized.