1: \begin{abstract} 

2:  This work deals with the simulation of Wishart processes and affine

3:  diffusions on positive semidefinite matrices. To do so, we focus on the

4:  splitting of the infinitesimal generator, in order to use composition

5:  techniques as Ninomiya and Victoir~\cite{NV} or Alfonsi~\cite{Alfonsi}. Doing so, we

6:  have found a remarkable splitting for Wishart  processes that enables us to

7:  sample exactly Wishart distributions, without any restriction on the

8:  parameters. It is related but extends existing exact simulation methods based on

9:  Bartlett's decomposition.  Moreover, we can construct high-order discretization

10:  schemes for Wishart processes and second-order schemes for general affine diffusions.

11:  These schemes are in practice faster than the exact simulation to sample

12:  entire paths. Numerical results on their convergence are given.

13: %%  At the same time, we present a

14: %%  general recursive method, based on the idea studied by Ninomiya and Victoir, for

15: %%  getting the second weak order scheme for any affine processes defined

16: %%  on the symmetric positive cone matrices. Therefore, these results allow us to

17: %%  propose a second order scheme for more general affine diffusion

18: %%  involving affine positive matrices. Simulation examples will be given

19: %%  to emphasize the convergence of schemes on Wishart process. We also

20: %%  give an application in finance for the model presented by Gourieroux

21: %%  and Sufana~\cite{Gourieroux} which can be seen as an extension of the

22: %%  Heston model in multi dimension.

23: \end{abstract}