An improved algorithm for convex optimal uncertainty quantification with polytopic canonical form

In general, we consider optimal uncertainty quantification (OUQ) framework to address uncertainty factors in the absence of sufficient knowledge on them. An important class of optimal OUQ problem that can be solved via convex optimization methods had been studied by researchers. If the object function can be rewritten in polytopic canonical form (PCF), the optimization problem can be solved by exact or approximate iterative algorithms easily. However, there exist some weaknesses for the two algorithms in computation demanding and computation accuracy. In order to overcome these weaknesses, we propose a new approximate iterative algorithm considering contribution of every constraint and verify the effectiveness of the new algorithm by simulation results.