We have developed a g-inverse method for solving geometric programming problems with negative degrees of difficulty. Such a problem is said not to have solution, but recently very few researchers came up with some method for solving such problem which we see as yielding infeasible optimal solution. We point out shortcomings in the approach and developed a better method that achieves dual feasibility and global optimal solution.

operations research, mathematical programming, nonlinear programming, geometric programming, negative degree of difficulty, global optimal solution