A new approach to uncertain parameter linear programming (Q1179011)

A new method for solving linear programming problems is given where the parameters are subject to uncertainty. No assumptions are made about probability distributions, but the method depends on specifying upper and lower bounds for each parameter. Solutions are obtained by solving successive deterministic linear programming problems. Convergence of the algorithm is proved. A definition and discussion of decision risk for this type of problem is given. There is a brief discussion of existing stochastic linear programming and fuzzy linear programming procedures which depend on assumptions about probability distributions of the parameters. The new method is applied to an example of feeding various livestock on various feeds.