A fast convergence points set generation module for high-dimensional pricing problems.

For pricing purpose, Monte-Carlo is slow but deterministic methods are unsuitable for long maturity products. This dead-end paves the way to new hybrid techniques mixing randomness and quadrature rules for simulating efficiently the diffusion of financial assets.

Features

Fast convergence

Accelerated Monte-Carlo Component™ guarantees a convergence rate function better than Monte-Carlo even for high-dimensional pricing problems (long maturity products, baskets with numerous underlyings).
The gain in computation time is significant compared to standard low-discrepancy sequences when the dimension becomes bigger than 50. For small dimensions, it behaves like the best known quadrature rules.
The running-time improvement is overwhelming compared to basic Monte-Carlo as soon as a significant accuracy is required.

Reliable precision estimation

Traditionnal low-discrepancy sequences-based methods do not provide any estimator of the accuracy of your pricing. Then when should you stop the simulation ?
Accelerated Monte-Carlo Component™ integrates a theoretically-proven precision estimator which tells you at each iteration of the simulation which accuracy you have reached. This way you do not lose the reliability of the traditionnal Monte-Carlo method.

Optimal implementation

The design of the software component relies on advanced precomputation schemes to stay as close as possible from what the theory predicts. The implementation uses numerical shortcuts written in assembly to offer you the best performances you can expect from your hardware.

Easy integration

Accelerated Monte-Carlo Component™ is packaged in a well-designed and fully documented API available in the C and C++ languages which ensures a simple and direct integration of the module in your pricing application. The module is available for Windows 95/98/NT/2000/XP and Unix/Linux platforms.

Benefits

• Fast convergence for high-dimensional pricing superseding both Monte-Carlo and Quasi-Monte-Carlo methods
• Combines the advantages of both Monte-Carlo and Quasi-Monte-Carlo methods
• Reliable precision estimation
• Tight implementation with low-level speed-up's for optimal performances
• Easy integration