Nonlinear superprocesses (Q2563931)

A nonlinear superprocess in the McKean-Vlasov sense is defined as a superprocess \((X_t)_{t\geq 0}\) for which the generator \(L\) of the one-particle motion, the immigration function \(b\), and the branching rate \(c\) may depend on the flow of marginal distributions \(u_t=P \circ X_t^{-1}\) of the nonlinear superprocess itself. I.e., the process \(X_t(f)-X_0(f)- \int^t_0 X_s(L(u_s)f+ b(u_s)f)ds\) is a local martingale with respect to \(P\) and possesses the increasing process \(\int^t_0 X_s(c(u_s)f^2)ds\), for each function \(f\) taken from the common domain of all generators \(L(m)\), \(m\) being a probability measure on the state space of \(X\). The paper gives Lipschitz conditions on the characteristics \(L,b\), and \(c\) in terms of an \(L^2\)-Wasserstein metric such that the above martingale problem has a unique solution. The proof uses stochastic calculus on historical trees.