A diffusion approximation in the ruin problem for a controlled Markov chain
This paper deals with the relationship between two-dimensional parameter Gaussian random fields verifying a particular Markov property and the solutions of stochastic differential equations. In the non Gaussian case some diffusion conditions are introduced, obtaining a backward equation for the evolution of transition probability functions.
The paper deals with the optimal inspections and maintenance problem with costly information for a Markov process with positive discount factor. The associated dynamic programming equation is a quasi-variational inequality with first order differential terms. In this paper we study its different formulations: strong, visousity and evolutionary. The case of impulsive control of purely jump Markov processes is studied as a special case.
Let be a fixed point of a substitution on the alphabet and let and . We give a complete classification of the substitutions according to whether the sequence of matrices is bounded or unbounded. This corresponds to the boundedness or unboundedness of the oriented walks generated by the substitutions.