Integral question

Soldato
OP
Joined
2 Jul 2010
Posts
3,098
It's the integral of the final term, ie. after the second plus sign. Recall that the integral of e^-at is (1/a)(1-e^-at)

I have that bit but if you take the integral of the existing integral it's not as simple as that. In this cause we have e^-a(T-u) and so if you integrate that between T and 0 you won't end up with it an equation in the form you mention above.

I also cannot understand how it's gone from integrating between 0 and t with respect to Wu to integrating between 0 and T with respecting to Wu.

I'm not sure if I am failing to apply a really fundamental rule of calculus or if there is some result that is very specific to reaching the above integral that I am not aware of.
 
Associate
Joined
7 Dec 2002
Posts
2,188
H7RPmpE.png

Can somebody please explain to me how the bit in yellow was obtained? I can't figure it out.

For clarity, Wu is a Wiener process.

In the first line extend the upper limit of the integral to infinity and replace the integrand with exp[-a(t-u)]H(t-u), where H is the Heaviside function. Then integrate your expression and interchange the order of integration on the final term (valid application of Fubini's theorem since the integrand is absolutely integrable with respect to the product measure dt dW(u)). This will yield the result after a couple of lines of computation.
 
Soldato
Joined
4 Feb 2003
Posts
6,118
Location
Birmingham
In the first line extend the upper limit of the integral to infinity and replace the integrand with exp[-a(t-u)]H(t-u), where H is the Heaviside function. Then integrate your expression and interchange the order of integration on the final term (valid application of Fubini's theorem since the integrand is absolutely integrable with respect to the product measure dt dW(u)). This will yield the result after a couple of lines of computation.


Exit for GD is that way --->

Get out and do not come back until you have read the unwritten rules and can recite the exact dimensions of all the Mods letterboxes :mad:
 
Soldato
OP
Joined
2 Jul 2010
Posts
3,098
In the first line extend the upper limit of the integral to infinity and replace the integrand with exp[-a(t-u)]H(t-u), where H is the Heaviside function. Then integrate your expression and interchange the order of integration on the final term (valid application of Fubini's theorem since the integrand is absolutely integrable with respect to the product measure dt dW(u)). This will yield the result after a couple of lines of computation.

Thanks for a serious answer!

I managed to solve this by using a different method in the end. Ultimately, I was making a mistake when changing the limits whilst changing the order of integration. By putting that step right, I was able to get the desired result.
 
Soldato
OP
Joined
2 Jul 2010
Posts
3,098
Potato.

But what does it mean :confused:

It's an interest rate model and by using the above result you can price bonds and other interest rate derivatives with it. However, it is too simplistic (compared to other models at least) to be used in practice :p.
 
Caporegime
Joined
17 Feb 2006
Posts
29,263
Location
Cornwall
I did A level maths in 6th form ( im 44 now ) and 1 look at all that and i realize that i am not able to remember a damn thing............. 3 years wasted
I did A-level maths. I failed it :p

Looking back I would not have used a single thing I learned (or failed to learn) in either a professional or personal capacity.

But yeah, maths is hard.
 
Associate
Joined
17 Oct 2009
Posts
2,348
What's it from?

I did A level maths in 6th form ( im 44 now ) and 1 look at all that and i realize that i am not able to remember a damn thing............. 3 years wasted
I did degree level maths, and I'm younger than ya, I also can't remember it! Sad times :(
 
Soldato
Joined
7 Apr 2009
Posts
6,833
Location
Western Seaboard
VpHeMSV
VpHeMSV
VpHeMSV
W8puUqq.jpg
 
Last edited:
Back
Top Bottom