Handling Contradicting Requirements Using Desirability Functions

Desirability functions serve to turn multiple response optimization problems into single response problems. A new set of these functions will be introduced within this paper that are based on an analogous probabilistic model for predicting the optimal starting time of a process to meet a given time schedule. We use loss-function-like functions and the probability of not meeting a given target to find the optimal set of input parameters where statistical properties are as we expect.