Invertible function y=f(x), x=f^(-1)(y) with two linear segments and smooth transition

Click For Summary
SUMMARY

The discussion focuses on constructing an invertible function y=f(x) that transitions smoothly between two linear segments, Ax+B for x<>T, with T being the crossover point. The proposed solution involves using sigmoid-like functions or peak-shaped functions integrated twice to achieve the desired transition. Additionally, a piecewise definition of the inverse function f^(-1) is suggested, utilizing cubic splines for smooth connections between segments. The challenge remains in ensuring the overall function is invertible, particularly when using simpler forms like y=ln(exp(Ax+B) + exp(Cx+D).

PREREQUISITES
  • Understanding of piecewise functions and their properties
  • Familiarity with sigmoid functions and their applications
  • Knowledge of cubic splines for smooth transitions
  • Experience with computer algebra systems, specifically Maxima
NEXT STEPS
  • Research the properties of invertible functions in mathematical analysis
  • Explore the application of cubic splines in function approximation
  • Investigate various sigmoid functions and their invertibility
  • Learn about piecewise function definitions and their implications in calculus
USEFUL FOR

Mathematicians, engineers, and researchers working on force-displacement relationships in materials, particularly those dealing with cracked solids and requiring smooth transitions in piecewise functions.

smartscience
Messages
1
Reaction score
0
Hi,

I'm looking to find a function y=f(x), invertible to x=f(y) and written in terms of elementary functions and operations, that can represent a straight line Ax+B where x<<T and another straight line Cx+D where x>>T, where T is the x position where the two lines would cross. In the region where x is approximately T, there should be a smooth transition between the two functions. For what it's worth, my intention is to get a functional representation of the force-displacement relationship for a cracked solid, which becomes stiffer when the crack is closed.

Some suitable functions can be found by integrating any sigmoid-like function, or equivalently integrating any peak-shaped function twice. The width of the peak or sigmoid then becomes the width of the transition from one line to the other. Similarly, a sigmoid function could be used as a weighting factor for terms Ax+B and Cx+D, e.g.

y=(Ax+B)(1/(1+exp(x))) + (Cx+D)(1/(1+exp(-x)))

Simpler functions like

y=ln(exp(Ax+B) + exp(Cx+D))

also spring to mind. However, these functions don't appear to be invertible either with my mathematical knowledge as it stands, nor with the computer algebra package Maxima.

Does anyone know any solutions to this kind of problem please? If this is a well-known problem, some terms to search for on would be appreciated. If there's no obvious solution, I'd even welcome just some suggestions for sigmoid functions that are more likely to yield an invertible function when used in this way.

Thanks in anticipation, Joe
 
Mathematics news on Phys.org
Very late response, but in case someone else is interested, this is what I would do. Instead of dividing the interval into two pieces, around x = T, I would divide it into three pieces: ##-\infty < A < T < B < \infty##.
The inverse, ##f^{-1}## could be defined in a piecewise fashion on the three separate intervals ##(-\infty, A), (A, B), \text{ and } (B, \infty)##. On the left interval you could define the inverse as y = ax + b, and on the right interval as y = cx + d. Notice that a and A etc. are different numbers. A cubic spline could be used to provide a smooth connection between the two outer parts of the graph.
 

Similar threads

  • · Replies 2 ·
Replies
2
Views
1K
Replies
3
Views
2K
  • · Replies 13 ·
Replies
13
Views
3K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 3 ·
Replies
3
Views
1K
  • · Replies 4 ·
Replies
4
Views
4K
Replies
8
Views
2K
  • · Replies 12 ·
Replies
12
Views
3K
Replies
17
Views
3K