Comparison of stochastic Volterra equations (Q5933566)

The authors prove a pathwise comparison theorem for solutions of the following one-dimensional stochastic Volterra equations: \[ X_i(t)=\xi_i + \sum_{j=1}^n H(t)\int_0^t \sigma_j(s,X_i(s)) dM_j(s) + \sum_{j=1}^n \int_0^t b_j^i(t,s,X_i(s)) dV_j(s),\quad i=1,2, \] where \(\{H(t)\}\) is a continuous adapted positive and strictly decreasing process, \(\{M_j(t)\}\) are continuous local martingales, and \(\{V_j(t)\}\) are continuous processes with bounded variation. The main assumptions are: \(\xi_1\leq\xi_2\), \(b_j^2(s,r,y)\geq b_j^1(s,r,x)\) for \(0\leq r\leq s\) and \(x \leq y\), \(b_j^2(s',r,y)-b_j^1(s',r,x)\geq b_j^2(s,r,y)-b_j^1(s,r,x)\) for \(0\leq r\leq s \leq s'\) and \(x \leq y\), and the statement has the form \(P(X_1(t) \leq X_2(t) \;\forall t)=1\).