Duan Markov Chain for Discrete Barrier options

Royboy · August 9th, 2006, 7:34 am

Hi,I am trying to code the Duan paper to value discrete barrier options, but I dont quite understand how to implement the probability transition matrix.Could anybody give me any hints or tips as to how to code this paper?Thank You

spursfan · August 10th, 2006, 6:07 am

RoysanI though the whole idea of the summer dissertation was that you did most of the work yourself (and yes I did notice my QUAD code that you're trying to convert into Matlab) - I wouldn't bother with Duan, the QUAD method should be betterMike

Royboy · August 10th, 2006, 8:29 am

I have been told by my supervisor (Dr Ales Cerny) that the Duan Markov chain method should be my benchmark. This method is better than Quadrature, since it is simplistic to implement and it uses the whole underlying asset price distribution instead of approximating the distribution using a polynomial.So which method do you suggest as the benchmark? (Maybe Monte Carlo?)Kind Regards

spursfan · August 10th, 2006, 2:50 pm

under the BS world, there's no need to use MC for benchmark barrier values with discrete samplingbroadie and yamamoto show values using Reiner's convolution method to 10dp

risingboy · June 6th, 2008, 11:03 am

Yes, it is too late to reply. But I hope it is helpful for others who has the similar problem.I implemented Duan's Markov Chain method for discrte barrier option. You may download the code and report here:http://www.mathworks.com/matlabcentral/ ... Type=fileI found this method was very fast and accurate. It is definitely a killer to the popular multilattice method.Ying

ckck20 · June 6th, 2008, 12:20 pm

Thanks, i'll have a look at it later on.