Evaluation of a horrible Multiple Integral

  • Context: Graduate 
  • Thread starter Thread starter Korbid
  • Start date Start date
  • Tags Tags
    Integral Multiple
Click For Summary

Discussion Overview

The discussion revolves around the numerical evaluation of a complex 8-dimensional integral related to the second virial coefficient in statistical mechanics. Participants explore methods to simplify the integral, which involves variables for position and velocity of two particles.

Discussion Character

  • Exploratory, Technical explanation, Mathematical reasoning

Main Points Raised

  • One participant seeks to simplify the integral and expresses concern about its complexity.
  • Another participant suggests a change of variables to reduce the dimensionality of the integral, proposing to let \(\vec{r} = \vec{r_{12}}\) and \(\vec{v} = \vec{v_{12}}\).
  • A participant challenges the initial suggestion, noting that the variable \(\vec{v_2}\) remains in the first exponential after the proposed change.
  • In response, a new change of variables is proposed, letting \(\vec{v} = \vec{v_2} - \vec{v_1}\) and \(\vec{u} = \vec{v_1} + \vec{v_2}\), which leads to a new expression for the sum of the squares of the velocities.
  • One participant reports success in simplifying the integral further, providing a new integral expression with updated variables.

Areas of Agreement / Disagreement

Participants do not reach a consensus on the best method for simplification, as there are differing approaches and some confusion regarding the implications of the proposed changes of variables.

Contextual Notes

Participants express uncertainty about the effectiveness of the proposed simplifications and the remaining complexity of the integral. The discussion reflects ongoing exploration rather than definitive solutions.

Korbid
Messages
15
Reaction score
0
I need to evaluate numerically this integral (second virial coefficient).
At first, i want to simplify a bit, how could i do it?
Now, i have a 8-dimensional integral and it's very horrible.
T, k, tau0 and R are constants.
r is the position and v is the velocity for two particles 1,2.
r12 = r1-r2; v12=v1-v2

$$\int{e^{-\frac{(|\vec{v}_1|^2+|\vec{v}_2|^2)}{2T}}e^{-E(\tau)/T}}d^2r_1d^2r_2d^2v_1d^2v_2$$
$$E(\tau)=\frac{k}{\tau^2}e^{-\tau/\tau_0}$$
$$\tau(\vec{r}_{12},\vec{v}_{12})=\frac{-(\vec{r}_{12}\cdot\vec{v}_{12}) -\sqrt{(\vec{r}_{12}\cdot\vec{v}_{12})^2-|\vec{v}_{12}|^2(|\vec{r}_{12}|^2 - (2R)^2)}}{|\vec{v}_{12}|^2}$$
 
Physics news on Phys.org
Well, you can simplify it a little by changing variables: Let \vec{r} = \vec{r_{12}} and \vec{v} = \vec{v_{12}}. Then you can integrate over \vec{r_1} and \vec{v_1} to leave just an integral over \vec{r} and \vec{v}.
 
stevendaryl said:
Well, you can simplify it a little by changing variables: Let \vec{r} = \vec{r_{12}} and \vec{v} = \vec{v_{12}}. Then you can integrate over \vec{r_1} and \vec{v_1} to leave just an integral over \vec{r} and \vec{v}.

stevendaryl,

At first, thanks for answering me quickly.

I'm sorry but i don't understand your method.

if we do this change then we still have \vec{v2} in the first exponential.
 
Korbid said:
stevendaryl,

At first, thanks for answering me quickly.

I'm sorry but i don't understand your method.

if we do this change then we still have \vec{v2} in the first exponential.

Okay, so instead, let's let \vec{v} = \vec{v_2} - \vec{v_1} and \vec{u} = \vec{v_1} + \vec{v_2}. Then |v_1|^2 + |v_2|^2 = \frac{1}{2}(|\vec{u}|^2 + |\vec{v}|^2)

Then you can do the integral over \vec{u}.
 
  • Like
Likes   Reactions: Korbid
stevendaryl said:
Okay, so instead, let's let \vec{v} = \vec{v_2} - \vec{v_1} and \vec{u} = \vec{v_1} + \vec{v_2}. Then |v_1|^2 + |v_2|^2 = \frac{1}{2}(|\vec{u}|^2 + |\vec{v}|^2)

Then you can do the integral over \vec{u}.

Well, i have reached to simplify enough the integral. Thank you.
$$\int^{\infty}_0\int^{\infty}_0\int^{2\pi}_0 rv\exp\left(-\frac{v^2}{4T}\right)\exp\left(-\frac{E(\tau)}{T}\right)\;\mathrm{d}r\mathrm{d}v\mathrm{d}{\theta}$$
$$E(\tau)=\frac{k}{\tau^2}\exp\left(-\frac{\tau}{\tau_0}\right)$$
$$\tau=\frac{-r\cos(\theta)-\sqrt{(2R)^2-r^2\sin^2(\theta)}}{v}$$
 

Similar threads

  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 7 ·
Replies
7
Views
887
Replies
3
Views
1K
Replies
1
Views
2K
  • · Replies 2 ·
Replies
2
Views
3K
  • · Replies 8 ·
Replies
8
Views
2K
Replies
2
Views
1K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 10 ·
Replies
10
Views
3K