Rubel’s universal differential equation

The starting point, g(x), is a rubelequation_1.gif function which vanishes outside the interval (-1,+1) :

rubelequation_2.gif

rubelequation_3.gif

rubelequation_4.gif

Its primitive is constant outside the interval (-1,+1), resp. f(-1)=0 if x<-1 and f(1)=Ω if x>1 :

rubelequation_5.gif

rubelequation_6.gif

rubelequation_7.gif

Ω is a well-defined real (no closed form is known) :

rubelequation_8.gif

rubelequation_9.gif

Rubel’s model, f(x), fits the given function y(x) in the interval (a,b) (exactly at the ends of the interval) :

rubelequation_10.gif

Example of a continuous junction between two consecutive intervals, say (0,1) and (1,3/2) (numerical integration is needed) :

rubelequation_11.gif

rubelequation_12.gif

rubelequation_13.gif

Differential equation fullfilled by the model function, f(x), whatever the interval :

rubelequation_14.gif

rubelequation_15.gif

rubelequation_16.gif

rubelequation_17.gif

Set W=rubelequation_18.gif and  eliminate W, a, b  and s[x]  between the five equations :

rubelequation_19.gif

rubelequation_20.gif

Mathematica needs a little help.  Obviously, the two first equations allow to eliminate W and (a-b) respectively :

rubelequation_21.gif

rubelequation_22.gif

rubelequation_23.gif

rubelequation_24.gif

This is Rubel’s equation where fk stands for the kth derivative of y(x).

Spikey Created with Wolfram Mathematica 8.0