You are on page 1of 3

J. M.

Williams

Market Growth ODE Suggestion

John Michael Williams, Ph. D. jmmwill@comcast.net [jwill@BasicISP.net] 1996 December 30 [reformatted 2012-05-07] To The Editor Engineering Management Review Box 5740 Clarkson University Potsdam, NY 13699-5740 Dear Dr. Wells: I wanted to congratulate you for the very informative and enjoyable Winter 1996 issue. I wonder if it was more than a coincidence that your preface began with some calculus? I also have a few comments on the differential equation representing market acceleration = volumetric growth * market headroom:

d 2V dV c1 * ( peak c 2V ). V dt dt 2
First, unless I am misunderstanding your intention, the Vpeak term just represents the value at the maximum time under consideration. But, because the volume is normalized to be 1 at maximum time 1, the Vpeak might as well just be 1. Along these same lines, because solution of the differential equation itself would introduce two or three free parameters, there really should be no need for the (I infer) arbitrary free parameters c1 and c2. With enough free parameters ANYTHING can be made to fit anything else! So, I'd see no loss of generality in restating the problem as:

d 2V dV * ( V) 1 2 dt dt

J. M. Williams

Market Growth ODE Suggestion

It's interesting to speculate why this differential equation would be better than some other approach. I think the only obvious reason would be if the market acceleration were easier to measure than some other characteristic. So, I would assume that one would use the differential equation by measuring market acceleration over a period of time, recording, perhaps, volumetric growth and current market volume. I guess that because both market acceleration and volumetric growth almost surely would have to be derived from the successive measurements of current market volume, the solution(s) of the differential equation would be far more important than the equation itself. This again just reinforces the idea of not parameterizing the differential equation before solving it. My own approach to the use of the differential equation you gave, would be to look at the scaling of time and volume: The problem actually is not one of time, but of shape of a curve on an axis which happens to have been derived from a time axis. To allow the problem to fit the solution, I would state it this way:

d 2V dV * ( V ), 1 dt dt 2
constrained by the boundary conditions V(0) = 0 and V(1) = 1. However, staying in the problem space just a little longer, I notice that the differential equation can be understood by looking at its two, right-hand side terms, separately:

d 2V dV * (misc) 2 dt dt
d 2V ( V ) * (misc) 1 dt 2

J. M. Williams

Market Growth ODE Suggestion

The second equation has a general solution of a sum of sine and cosine functions, each scaled by an arbitrary constant. Choosing V(t) = sin(t) - 3*cos(t) produces a nice, S-shaped curve:

In the actual differential equation, my first equation just above would represent a scaling factor giving another couple of degrees of freedom in the precise shape of the solution. Do you think the volumetric growth term is necessary?

Yours truly, [John Michael Williams]

You might also like