The communication presents results concerned optimal stopping algorithms for American type options for multivariate modulated Markov type price processes with discrete and continuous time. New space and time skeleton approximation algorithms, tree type backward recurrence approximation algorithms, and Monte Carlo based algorithms are described as well as the corresponding convergence results. Applications to models of exchange of assets, reselling of options and optimal option pricing for mean-reverse models used to describe stochastic dynamics of energy prices are given and illustrated by results of experimental studies.
Keywords: Optimal stopping; American option; Approximation algorithm; Reselling of options
Biography: Dmitrii Silvestrov is a professor in mathematical statistics at Stockholm University. He has degree of Dr.Sc and is a member of the International Statistical Institute, author of 8 books and more that 120 research papers in the areas of stochastic processes, financial and actuarial mathematics, and statistical software.