Postoptimality Analyses

Abstract

As discussed in Chapter 1 of this volume, one of the assumptions in linear programming is that all parameters are deterministic, i.e., assumed to be known with certainty. Clearly, this assumption will rarely be satisfied in practice. One popular way to get around the problem, i.e., dealing with uncertainty while keeping the simple structure and efficient solution methods of linear programming, are sensitivity analyses. In essence, sensitivity (sometimes also called “what...if?”) analyses deal with the effects of parameter changes on the optimal solution. Typical examples include analyses regarding the effects of price changes on a cutting plan, changes in the demand structure on the allocation of resources, and technological changes (such as faster throughputs in some machines) on a production schedule.