Some experimental results to illustrate the performance of randomized approximation algorithm and its derandomized version, are presented. The quality of the option price computed by algorithms are compared with those by other approximation algorithms. The full-path method is implemented to compute the exact price, and approximation algorithms such as the naive Monte Carlo method (MC), the AMO algorithms (AMO-LB, AMO-UB), the DHL algorithms, and the OSST algorithm. A naive method for computing the exact price of European-Asian options, called the full-path method, enumerates all paths in the binomial tree model. The full path method requires an exponential time since there are exponential number of paths in the binomial tree. The error bound of the Mount Carlo method depends volatility of the stock price when a polynomial number of samples are taken by naive sampling.
ASJC Scopus subject areas
- コンピュータ サイエンスの応用