Approaches and Mathematical Models for Robust Solutions to Optimization Problems with Stochastic Problem Data Instances
Practical applications of scheduling, routing and other generic constrained optimization problems often involve an uncertainty in the values of the data presented in the problem data instances. On the contrary, most of the established algorithms for typical classes of well-studied problems in the field of constrained optimization assume that deterministic precise values of data would be known. Hence, any solution developed for a specific optimization problem with a given problem data instance would become non-optimal and/or infeasible when applied to another data instance with even slight perturbation. We argue the fallacy of using solutions developed based on the mean values of data for real life problems having stochastic data.
Unable to display preview. Download preview PDF.