A new proof for comparison theorems for stochastic differential inequalities with respect to semimartingales (Q1807276): Difference between revisions

The two systems of stochastic integral inequalities \[ X_i(t)\leq X_i(s)+\int_s^ta_i(u,X(u))dA_i(u)+\sum_{k=1}^r \int_s^t\sigma_{ik}(u,X(u))dM_k(u),\tag{1} \] \[ Y_i(t)\geq Y_i(s)+\int_s^tb_i(u,Y(u))dA_i(u)+\sum_{k=1}^r \int_s^t\sigma_{ik}(u,Y(u))dM_k(u) \tag{2} \] are concerned, where \(0\leq s\leq t\), \(i=1,2,\dots,d\), \(M=(M_1,\dots,M_r)\) is a local martingale, \(A_i\) are one-dimensional continuous increasing processes and the (random) coefficients \(f_i,a_i,b_i,\sigma_{ik}:\mathbb R_+\times \Omega \times \mathbb R^d\to \mathbb R\) are \(\mathcal P\otimes \mathcal B(\mathbb R^d)\)-measurable (where \(\mathcal P\) denotes the predictable \(\sigma \)-field) and pathwise continuous. Under suitable conditions comparison theorems are proved for the systems of inequalities (1), (2). Furthermore, the maximal/minimal solutions are constructed using the monotone iterative technique. As an application a stochastic Bihari-type inequality is derived in one-dimensional case.