Homeomorphic flows for multi-dimensional SDEs with non-Lipschitz coefficients (Q2485835): Difference between revisions

Consider the multi-dimensional stochastic differential equation (SDE): \[ dX_t =\sigma(t, X_t)\, dw_t + b(t, X_t)\, dt,\quad X_0=x\in\mathbb R^d,\tag{1} \] where \(\sigma : R_+\times\mathbb R^d\mapsto \mathbb R^d\times\mathbb R^m\), \(b : R_+ \times\mathbb R^d\mapsto\mathbb R^d\) are Borel measurable functions: \[ \|\sigma(t,x)-\sigma(t,y)\|\leqslant\| x-y\|^2\cdot g_1(\| x-y\|),\tag{2} \] \[ \| b(t,x)-b(t,y)\|\leqslant\| x-y\|\cdot g_2(\| x-y\|),\tag{3} \] for all \((t,x,y)\in\mathbb R_+\times\mathbb R^d\times\mathbb R^d\), where \(g_i\) are positive continuous functions, bounded on \([1,+\infty)\) and satisfying \[ \lim_{x\downarrow0} \frac{g_i(x)}{\log x^{-1}}=\delta_i<\infty,\quad i=1,2.\tag{4} \] Here \(w_t\) is an \(m\)-dimensional standard Brownian motion defined on classical Wiener space \((\Omega,\mathcal F,P)\), namely \(\Omega = C_0(\mathbb R_+, \mathbb R^d)\) with uniform convergence topology, \(P\) is the Wiener measure and \(\mathcal F\) is the \(P\)-completion of Borel field \(\mathcal B(\Omega)\). The stochastic integral is understood as Itô's integral. The author studies the homeomorphic flow property of solutions of SDE (1) with non-Lipschitz coefficients (2)--(4) for small \(t\), which is then extended to arbitrarily large \(t\). Zvonkin's transformation is applied to weaken the condition imposed on the drift term \(b\).