Exact simulation of jump-diffusion processes with Monte Carlo applications (Q660166): Difference between revisions
From MaRDI portal
Added link to MaRDI item. |
Removed claim: reviewed by (P1447): Item:Q456198 |
||
Property / reviewed by | |||
Property / reviewed by: Martin Georg Riedler / rank | |||
Revision as of 02:32, 15 February 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Exact simulation of jump-diffusion processes with Monte Carlo applications |
scientific article |
Statements
Exact simulation of jump-diffusion processes with Monte Carlo applications (English)
0 references
26 January 2012
0 references
The authors introduce an exact simulation algorithm (JEA) for one-dimensional diffusion processes with jumps possessing state-dependent intensities. Here, `exact' means that the algorithm allows for unbiased Monte Carlo estimation of (path-dependent) functionals. The present work is an extension of the Exact Algorithm (EA) for diffusions (without jumps) [\textit{A. Beskos, O. Papaspiliopoulos} and \textit{G. O. Roberts}, Bernoulli 12, No. 6, 1077--1098 (2006; Zbl 1129.60073)]. More precisely, the method is valid for stochastic differential equations (SDEs) of the form \[ dX_t=\alpha(X_{t-})dt+dW_t+\int_E \Delta(z,X_{t-})\,m(dz,dt),\quad X_0=x_0, \] where \(m\) is random point measure with state dependent intensity \(\lambda(X_{t-})\). The diffusion coefficient \(\alpha\) needs to satisfy the same condition as for the EA in the case without jumps and the jump intensity \(\lambda\) has to be bounded. The last condition is necessary to employ a thinning argument which transforms the problem of state dependent intensities to constant intensities with the effect of resulting in a discontinuous jump coefficient. This allows the exact simulation of the jump times of the resulting equation. Though the JEA is restricted to the above type of SDEs an inverse transformation allows for most general one-dimensional SDEs satisfying usual Lipschitz and linear growth conditions to be transformed to the above form. Often the inverse transformation is possible without biasing the Monte Carlo estimators, thus the algorithm is exact also for the transformed equation. Analogously to the EA, the essential idea is reducing questions on jump diffusion to questions on the Brownian Bridge via the EA's skeleton construction. Thus, if the problem of simulating a functional is solved for the Brownian Bridge it can also be done for the jump diffusion. Finally, applications of the JEA to pricing of financial products are presented. The authors discuss the Monte Carlo simulation of functionals depending on the path at a finite set of fixed times and also path-dependent functionals over a finite time horizon. A particular example, called the SINUS model, is considered and the performance of the JEA is compared to the naive Euler scheme and the Euler scheme with thinning.
0 references
jump diffusion
0 references
simulation
0 references
exact algorithm
0 references
barrier option pricing
0 references