In applications of geometric programming, some coefficients and/or exponents may not be precisely known. We indicate which stochastic programming approaches and which structural and distributional assumptions do not destroy the favorable structure of geometric programs.
The already recognized possibilities are extended for a tracking model and stochastic sensitivity analysis is presented in the context of metal cutting optimization.