SciComp/NVIDIA

Forum Navigation:


FORUMS > Student Forum < refresh >
Topic Title: What is adaptedness of a stochastic process?
Created On Tue Apr 22, 03 09:54 PM
Topic View:

View thread in raw text format


asd
Senior Member

Posts: 660
Joined: Aug 2002

Tue Apr 22, 03 09:54 PM
User is offline


Can someone please give an intuitive explanation,by what is comfortable to be imagined when it is said that a process is f(t) adapted or f(t) measurable. Everytime I force the definitions in my mind,they pop-out and it is blank!
I have read it in every page that a process is f(t) adapted,but never read that it is not f(t) adapted - can there be cases where a process is not f(t) adapted ?

Can someone please explain it,or give a daily life example for intuition?


Thanks,
asd



-------------------------
Life is a journey and not a destination
quant code directory
 
Reply
   
Quote
   
Top
   
Bottom
     



Aaron
Senior Member

Posts: 6354
Joined: Jul 2001

Tue Apr 22, 03 10:44 PM
User is offline View users profile

"Measurable" and "adapted" refer to the process, not a realization. There are many different measures and process can be measurable (or not) with respect to, and many filtrations that a process can be adapted (or not) with respect to.

The simplest answer is these are technical regularity conditions that you don't have to worry about unless you explore untested realms of exotic models.

The next simplest answer is that these conditions allow you to do common sense things. If a process is measurable, then the probability distribution at any point in time can be integrated. It makes sense to talk about the probability that X(3)>12. If a process is adapted then you can answer questions about the path. For example, given the values of X(t) for 0>=t>=3, you can tell if x(t) was ever >12 in that interval.

An example of a non-adapted, non-measurable process is a random walk in which the variance at time t is equal to the lowest integer b such that t = a/b.

-------------------------
Aaron Brown
 
Reply
   
Quote
   
Top
   
Bottom
     



asd
Senior Member

Posts: 660
Joined: Aug 2002

Wed Apr 23, 03 02:53 AM
User is offline



Aaron,

Thank you very much for the insightful and excellent explanation and your help.


Regards,
asd





-------------------------
Life is a journey and not a destination
quant code directory
 
Reply
   
Quote
   
Top
   
Bottom
     



mj
Senior Member

Posts: 2930
Joined: Dec 2001

Wed Apr 23, 03 08:57 PM
User is offline View users profile

the basic problem of stochastic calculus is that the measure is defined on the space of paths which are functions of time. So when we draw a path at random, we simultaneously looking at all times.

However, being mere mortals we cannot see the future. We therefore need a concept that expresses
the information available at a given time. This concept is the filtration.

So F_t tells us what information there is at time t. If we say X_t is adapted to F_t then we are saying that F_t contains enough information to tell us what is true about X_t.

Often things are done in a backwards way in that we typically think of a Brownian motion and then define F_t to be the information needed to describe W_t and then we, of course, have that W_t is adapted to F_t.

When is something not adapted? If there is additional information that affects it. Eg if we have two different
Brownian motions W_t and W'_t with filtrations F_t and F'_t. then W_t is not adapted to F'_t.

We also have that W_{t+1} is not adapted to F_t as it tries to see the future.

MJ


-------------------------
LMM and QuantLib training course London June 2010. http://www.moneyscience.com/training/
 
Reply
   
Quote
   
Top
   
Bottom
     

View thread in raw text format
FORUMS > Student Forum < refresh >

Forum Navigation:

© All material, including contents and design, copyright Wilmott Electronic Media Limited - FuseTalk 4.01 © 1999-2010 FuseTalk Inc.