The theory of geometric programming is concerned with the solution of certain nonlinear programming problems in which the objective function and the constraints are polynomial expressions with positive coefficients. The solution of the primal geometric programming problem can be obtained by solving the dual problem. In certain cases the solution to the dual problem, and hence the solution to the primal problem, is obtained immediately by solving the dual constraint system for the optimal dual vector. Such problems are said to possess “degree of difficulty zero.” This paper describes a solution procedure for geometric programming. The procedure involves the formation of an augmented problem possessing degree of difficulty zero. The solution to the original problem requires the estimation of certain parameters in the augmented problem. An iterative procedure for estimating these parameters is described.