Large Deviations Of Semimartingales: A Maxingale Problem Approach.II. Uniqueness For The Maxingale Problem. Applications (Q4943565)

In part I [ibid. 61, No. 3/4, 141-243 (1997; Zbl 0890.60025)] the author reduced the problem of proving the large deviation principle (LDP) for semimartingales in a Skorokhod space to checking that a certain deterministic problem has a unique solution. Just this problem was called ``a maxingale problem''. The purpose of this paper is to establish conditions of uniqueness for the maxingale problem and to use these results to derive LDPs for stochastic processes. The author starts by briefly recalling the setting and notation of the first part and then states the uniqueness results. Section 2 gives more applicable versions of part I results and combines them with the above uniqueness results to derive LDPs for processes in a Skorokhod space. As in part I the author considers a sequence of \(R^d\)-valued semimartingales, defined on a common probability space and adapted to respective filtrations. The last section contains the proof of the mentioned uniqueness results. It is interesting to note that the obtained uniqueness conditions are similar in form to those for the martingale problem. Another fact worth mentioning is the important role played by the method of Orlicz spaces.