
Differential Equations Ordinary and Partial Differential Equations Math Forum 
 LinkBack  Thread Tools  Display Modes 
December 1st, 2016, 12:11 PM  #1 
Newbie Joined: Dec 2016 From: California Posts: 5 Thanks: 0  Whether to define system as PDE or ODE
Hey everyone. I'm a grad student revising a journal article. Within the article, I layout an equation which describes of population "$\displaystyle n(r)$" through time. The parameters of this population vary with some variable "$\displaystyle r$"; however, $\displaystyle r$ does not change with time. When I solve this system (written out below), I first solve for the "$\displaystyle N$value" at every value of r. Afterwards, I use an ODE solver to evaluate the time evolution for "$\displaystyle n(r)$" at every $\displaystyle r$value. $\displaystyle \frac{dn(r)}{dt} = \frac{\dot{D}}{D_{0}} \bigg( N(r)  n(r) \bigg)  n(r) \exp \bigg( \Delta E/k_B T \bigg) \frac{P(r) s}{P(r) + s}$ My question is this: is it strictly correct to describe this as an ordinary differential equation? Specifically, when I write the derivative $\displaystyle dn(r)/dt$, is this a mistake? A reviewer suggested that it is a mistake, writing "since you talk about a function of two variables, $\displaystyle n(r,t)$, the derivative should be written as $\displaystyle \partial n(r,t) / \partial t$." My reasoning in treating it as an ODE is that the solution to this equation requires differentiation with respect to $\displaystyle t$ only. Thanks in advance! Nathan 
December 1st, 2016, 12:16 PM  #2 
Senior Member Joined: Sep 2015 From: CA Posts: 617 Thanks: 330 
if the function is $n(r, t)$ where $n()$ truly depends on time as well as on $r$ then your reviewer is correct. On the other hand if the function is rather $n(r(t))$ where $n()$ happens to vary with time only because it's single parameter $r$ varies with time then I'd say your notation is fine. 
December 1st, 2016, 01:34 PM  #3 
Math Team Joined: Dec 2013 From: Colombia Posts: 6,354 Thanks: 2086 Math Focus: Mainly analysis and algebra 
There is also the consideration as to whether $r$ is a variable or a parameter. That is to say that if, for any given evolution of $n(t)$, the parameter $r$ remains fixed, then we have an equation in a single free variable which is therefore an ODE. However, if $r$ is free to change within an evolution of $n$, we have a function $n(r,t)$ of two free variables and thus we have a PDE. 
December 1st, 2016, 02:44 PM  #4 
Newbie Joined: Dec 2016 From: California Posts: 5 Thanks: 0 
Yes, v8archie, this was my initial reasoning for describing the equation as an ODE. For every evolution of $\displaystyle n(t)$, the $\displaystyle r$value is constant at all $\displaystyle t$values. In other words, the final goal is to describe how $\displaystyle n(t)$ evolves for every $\displaystyle r$value, but the $\displaystyle r$values are considered onebyone during differentiation. Last edited by hydronate; December 1st, 2016 at 02:56 PM. 
December 1st, 2016, 02:59 PM  #5 
Global Moderator Joined: Dec 2006 Posts: 16,225 Thanks: 1150 
What do you solve to obtain $N(r)$?

December 1st, 2016, 03:08 PM  #6 
Newbie Joined: Dec 2016 From: California Posts: 5 Thanks: 0 
This is the expression which is evaluated first to determine the $\displaystyle N$value associated with every $\displaystyle r$value (the N and rho values on the righthand side of the equation are constants): $\displaystyle N(r)dr = N 4 \pi r^2 \rho \exp \Bigg( \frac{4 \pi r^3}{3} \rho \Bigg) dr$ Last edited by hydronate; December 1st, 2016 at 03:11 PM. 
December 1st, 2016, 03:16 PM  #7 
Global Moderator Joined: Dec 2006 Posts: 16,225 Thanks: 1150 
Why the "dr" on both sides? Also, is that equation dimensionally correct?

December 1st, 2016, 04:39 PM  #8 
Newbie Joined: Dec 2016 From: California Posts: 5 Thanks: 0 
That equation has been in the literature for a while now. It is certainly dimensionally correct on the r.h.s. of the equation. The l.h.s. is reproduced correctly from the literature, but is difficult to justify! (Seems to be m^2) The physical meaning is this. There are different concentrations of $\displaystyle N$ ($\displaystyle m^{3}$) electrons and $\displaystyle \rho$ ($\displaystyle m^{3}$) holes. As you move radially outwards from an electron site, the probability that you will find a hole within a thin shell between distance $\displaystyle r$ and $\displaystyle r+dr$ is $\displaystyle 4 \pi r^2 \rho \exp \Bigg( \frac{4 \pi r^3}{3} \rho \Bigg) dr$. Assuming that the number of electrons is much less than the number of holes, the number of electrons with a nearest neighbor between $\displaystyle r$ and $\displaystyle r+dr$ is therefore $\displaystyle N(r)dr = N 4 \pi r^2 \rho \exp \Bigg( \frac{4 \pi r^3}{3} \rho \Bigg) dr$ 
December 1st, 2016, 05:45 PM  #9 
Math Team Joined: Dec 2013 From: Colombia Posts: 6,354 Thanks: 2086 Math Focus: Mainly analysis and algebra 
You might be interested in the profile of $N$ for different radii changes over time. In that case, the equation becomes a PDE.

December 2nd, 2016, 08:47 AM  #10 
Newbie Joined: Dec 2016 From: California Posts: 5 Thanks: 0 
I am interested in how the profile of $\displaystyle N(r)$ changes through time, that is true. However, I only ever evaluate over the same $\displaystyle r$values for all of my time evolutions.


Tags 
define, ode, pde, system 
Thread Tools  
Display Modes  

Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Define circle when:  Hakundin  Geometry  8  April 22nd, 2014 02:33 PM 
How to define the set of all primes?  etham  Applied Math  5  September 9th, 2013 04:06 PM 
Define a function z = f(x,y) by f(0,0) = 0 and otherwise  mattturner83  Real Analysis  0  April 24th, 2012 12:17 AM 
Define X  Snevk  Algebra  18  December 23rd, 2011 02:32 PM 
What should I define  Qer  Algebra  1  March 4th, 2010 01:53 PM 