Dynkin Game of Stochastic Differential Equations with Random Coefficients, and Associated Backward Stochastic Partial Differential Variational Inequality

Zhou Yang,Shanjian Tang
DOI: https://doi.org/10.1137/110850980
IF: 2.2
2013-01-01
SIAM Journal on Control and Optimization
Abstract:A Dynkin game is considered for stochastic differentialequations with random coefficients. We first apply Qiu and Tang's maximum principle for backwardstochastic partial differential equations to generalize a Krylovestimate for the distribution of a Markov process to that of anon-Markov process and establish a generalized Itô--Kunita--Wentzellformula which allows the test function as an Itôprocess to take values in a Sobolev space. We then prove theverification theorem that the Nash equilibrium point and the value ofthe Dynkin game are characterized by the strong solution of theassociated Hamilton--Jacobi--Bellman--Isaacs equation, which is currentlya backward stochastic partial differential variational inequality(BSPDVI) with two obstacles. Finally, we prove the existence anduniqueness result and a comparison theorem forstrong solution of the BSPDVI.
What problem does this paper attempt to address?