We give slight conditions for the existence and characterization of optimal solutions for the Markov decision problem with average cost under m constraints of the same kind in Borel actions and states spaces. The optimal policies are convex combination of at most m + 1 deterministic policies.
This paper presents an innovative tool and method that allow efficient innovation of shape and topology of virtual parts at both mesh and CAD levels using optimization methods. The method consists of automatic variations of shapes in CAD/CAE environments that allow effective search for new shapes that are not considered initially by designers.
The paper is concerned with the estimation of the change point of traffic intensity for a M/G/1 queue.