Difficult Differential Equations of Motion

**RFurball** · Jun26-09, 08:58 PM

I saw this problem posted on the internet somewhere and am intrigued how to solve it.
The problem is a fox chasing a hen with the following conditions:
1) The fox's starts at position x=0, y=10m (0,10)
2) The hen start's at position (0.0)
3) The fox runs at velocity 4m/s in a direction directly at the hen's current position.
4) The hen runs at a velocity of 3m/s along the x-axis.

The question is how long does it take the fox to catch the hen?

My approach to solve this starts with a position of the fox $LaTeX Code: F(t)=F_x(t) \\vec x+F_y(t) \\vec y$
At this positions the velocity vector is $LaTeX Code: VF(t)=VF_x(t) \\vec x+VF_y(t) \\vec y$ where

The position of the Hen is 3t
Based on the position of the 2 animals the ratio $LaTeX Code: \\frac{VF_x(t)}{VF_y(t)}=\\frac{3t-F_x(t)}{F_y(t)}$

Rearranging these equations I come up with the following 2 equations:

$LaTeX Code: VF_x(t)=\\frac{4(3t-F_x(t)}{\\sqrt{(3t-F_x(t))^2-F_y(t)^2}}$

and

$LaTeX Code: VF_y(t)=\\frac{4F_y(t}{\\sqrt{(3t-F_x(t))^2-F_y(t)^2}}$

From physics I know that $LaTeX Code: v(t)=\\frac{ds}{dt}$ which I think I can break into: $LaTeX Code: V_x(t)=\\frac{dF_x(t)}{dt}$ and $LaTeX Code: V_y(t)=\\frac{dF_y(t)}{dt}$

Plugging these 2 equations I come up with the following 2 equations which are a set of differential equations for Fx(t) and Fy(t):

$LaTeX Code: \\frac{dF_x(t)}{dt}=\\frac{4(3t-F_x(t)}{\\sqrt{(3t-F_x(t))^2-F_y(t)^2}}$
and
$LaTeX Code: \\frac{dF_y(t)}{dt}=\\frac{4F_y(t}{\\sqrt{(3t-F_x(t))^2-F_y(t)^2}}$

If I could solve for Fy(y) I would find t where Fy(t)=0

Are these equations solveable?
Secondly is my approach to solving correct?

This problem has got to be solveable since it seems simpler than the laws of planetary motion which involve accelaeration and 3 dimesions.

A coulple of observations I calculated are:
If the fox ran along the y-axis then along the x-axis the hen would be caught in 10 seconds so I know the aswer is less than 10seconds.
If the fox ran an angle of atan(3/7^0.5) the hen would be caught in 10/(7^0.5)=3.78seconds so I know the answer is greater than 3.78seconds.
If I run a recursive Excel spreadsheet with small delta t I calculate the answer to be approximately 5.715seconds.
I would use this number as a check of the solution.

Hope the post is legible. I am learning latex on the fly.

**D H** · Jun26-09, 10:08 PM

Hint: Denote F_x simply as x and F_y as y.

What is dy/dx?

**g_edgar** · Jun27-09, 07:28 AM

This problem (or equivalent) can be found in many differential equations textbooks:

Flying a plane in a crosswind, always heading straight toward the destination.

Rowing a boat across a river, always heading straight toward the destination.

I first heard it as a farmer chasing a pig.

Here is a solution I wrote up a few years ago

http://www.math.ohio-state.edu/~edga..._02/HW8s4.html

**D H** · Jun27-09, 08:13 AM

Don't give out solutions, g_edgar.

**g_edgar** · Jun27-09, 08:23 AM

Originally Posted by D H

Don't give out solutions, g_edgar.

OK, deleted.

**RFurball** · Jun27-09, 02:52 PM

DH

Thanks for your help. I am still a little confused where dy/dx comes into play, I will have to think about it. Maybe I will look in my 20yr old college text on differential equations to see if there is a similar probblem as g_edgar points out. I would guess the path is probably a parabola or hyperbola, maybe I can plug those into the equations.

**D H** · Jun27-09, 02:59 PM

THe path is called a tractrix, or pursuit curve.

Where dy/dx comes into play is that it is much easier to derive the curve y=f(x) rather than to derive x and y as functions of time. From that, you can easily compute the point at which this curve intersects the x axis. Finally, given the chicken's speed, you can easily compute the time it takes for the fox to catch the chicken.

**RFurball** · Jun28-09, 04:52 PM

D_H,

After thinking it over I think I understand now. The derivative dy/dx is the direction the fox is traveling. Therefore the solution would be solving this differential equation $LaTeX Code: \\frac {dy}{dx} = \\frac {-y}{3t-x}$ which I havn'e found the solution but that is OK. I just have bugged on how to solve it and I made it way more difficult than need be.

I did look in my college diff eq book and embarassing enough a similar problem is listed on page 12 (out of like 900+ total pages).

**clustro** · Jun29-09, 12:47 AM

For what its worth (not that it will assuage any pure mathematicians on this board), the problem was solved for all practical purposes when you made those two coupled ODE's.

They could be solved simultaneously by a variety of numerical methods.