Quadratic lower/upper bound of a function

Click For Summary
SUMMARY

The discussion focuses on finding second-order polynomial bounds for the function f(t) = exp(2t) within the interval 0.1 < t < 0.4. The user seeks to determine coefficients a, b, and c for the polynomial g(t) = a + bt + ct² such that g(t) < f(t) in the specified range. The user notes that f(t) does not need to be convex and expresses frustration over the lack of existing resources or systematic approaches for bounding functions of the form exp(αt), exp(αt)cos(βt), and exp(αt)cos²(βt).

PREREQUISITES
  • Understanding of polynomial functions and their properties
  • Familiarity with exponential functions and their behavior
  • Knowledge of calculus, particularly derivatives and bounds
  • Experience with mathematical optimization techniques
NEXT STEPS
  • Research methods for deriving polynomial approximations, such as Taylor series
  • Explore the topic of convex and non-convex functions in optimization
  • Investigate existing literature on bounding exponential functions
  • Learn about Chebyshev polynomials and their applications in approximation theory
USEFUL FOR

Mathematicians, researchers in optimization, and anyone involved in numerical analysis or approximation theory will benefit from this discussion.

aantam
Messages
2
Reaction score
0
Hi folks,

I have a function f(t), and I want to find 2nd order polynomials that lower/upper bound f(t) in a fixed interval. For instance,

f(t) = exp(2t), 0.1<t<0.4

Find a,b,c so that g(t) = a + b t +c t^2 <f(t) for the given interval

I have been googling for the solution, but apparently no one cares about this problem, although I was expecting it to be already solved :( Anyone could give me a reference to look at? Books, papers, whatever.. Oh, by the way, f(t) does not have to be convex.

Thanks a lot!
 
Physics news on Phys.org
Sorry, I forgot to mention, I assume some benign conditions for the function so that there always exist a lower and upper bound in the given interval (I could always pick a constant as the bounds but I want something better). Actually, the functions that I have to bound will always be of the form:

exp(\alpha t) , exp(\alpha t)*cos(\beta t), exp(\alpha t)*cos^2(\beta t)

It's weird that I didn't find a systematic approach for finding such bounds..

Thanks again
 

Similar threads

  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 3 ·
Replies
3
Views
4K
  • · Replies 5 ·
Replies
5
Views
3K
  • · Replies 2 ·
Replies
2
Views
4K
Replies
3
Views
2K
  • · Replies 4 ·
Replies
4
Views
3K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 23 ·
Replies
23
Views
4K
  • · Replies 13 ·
Replies
13
Views
3K