Abstract:We study a class of McKean--Vlasov Stochastic Differential Equations (MV-SDEs) with drifts and diffusions having super-linear growth in measure and space -- the maps have general polynomial form but also satisfy a certain monotonicity condition. The combination of the drift's super-linear growth in measure (by way of a convolution) and the super-linear growth in space and measure of the diffusion coefficient require novel technical elements in order to obtain the main results. We establish wellposedness, propagation of chaos (PoC), and under further assumptions on the model parameters we show an exponential ergodicity property alongside the existence of an invariant distribution. No differentiability or non-degeneracy conditions are required.
Further, we present a particle system based Euler-type split-step scheme (SSM) for the simulation of this type of MV-SDEs. The scheme attains, in stepsize, the strong error rate $1/2$ in the non-path-space root-mean-square error metric and we demonstrate the property of mean-square contraction. Our results are illustrated by numerical examples including: estimation of PoC rates across dimensions, preservation of periodic phase-space, and the observation that taming appears to be not a suitable method unless strong dissipativity is present.
What problem does this paper attempt to address?
### What problems does this paper attempt to solve?
This paper is mainly devoted to studying and solving problems related to McKean - Vlasov stochastic differential equations (MV - SDEs) with superlinear growth and their associated particle systems. Specifically, the author focuses on the following aspects:
1. **Well - posedness and Propagation of Chaos**:
- Firstly, the paper solves the well - posedness problem of MV - SDEs on the finite - time interval [0, T]. That is, it is proved that under given conditions, this type of equation has a unique strong solution.
- Secondly, the author explores the Propagation of Chaos (PoC), that is, when the number of particles tends to infinity, whether the empirical distribution of the system composed of N particles converges to the original MV - SDE.
2. **Exponential Ergodicity and Existence of Invariant Measure**:
- The author further studies the long - term behavior of MV - SDEs, especially its exponential ergodic property. This involves proving that the system has a unique invariant probability measure, and the solutions under all initial conditions converge to this invariant measure at an exponential rate.
- This property is very important in statistical inference, neuroscience, and the study of long - term behavior related to metastability.
3. **Numerical Approximation Scheme**:
- The paper proposes a numerical scheme based on the Split - Step Method (SSM) for simulating this type of MV - SDEs and associated particle systems.
- This numerical scheme achieves a strong convergence rate of order 1/2 under the non - path - space mean - square - error metric and shows the property of mean - square contraction.
### Main Contributions
- **Well - posedness and Propagation of Chaos**: The author introduces a new "additional symmetry" condition to overcome the technical difficulties brought by superlinear growth and proves the Lp moment estimate.
- **Exponential Ergodicity and Invariant Measure**: By analyzing the completeness in the space of probability measures, the exponential ergodic property and the existence of the invariant measure are proved without using Lyapunov functions or Krylov - Bogoliubov methods.
- **Numerical Method**: The proposed split - step method is not only suitable for simulation in a finite - time interval but can also be successfully applied to long - time simulation and shows good stability and robustness.
### Key Formulas
- **MV - SDE Form**:
\[
dX_t=\left(v(X_t,\mu_{X_t}) + b(t,X_t,\mu_{X_t})\right)dt+\sigma(t,X_t,\mu_{X_t})dW_t,\quad X_0\in L^m(\mathbb{R}^d),\;m > 2
\]
where,
\[
v(x,\mu)=\int_{\mathbb{R}^d}f(x - y)\mu(dy)+u(x,\mu),\quad\sigma(t,x,\mu)=\sigma(t,x,\mu)+\int_{\mathbb{R}^d}f_\sigma(x - y)\mu(dy)
\]
- **Wasserstein Distance**:
\[
W^{(r)}(\mu,\nu):=\inf_{\pi\in\Pi(\mu,\nu)}\left(\int_{\mathbb{R}^d\times\mathbb{R}^d}|x - y|^r\pi(dx,dy)\right)^{1/r},\quad\mu,\nu\in P_r(\mathbb{R}^