Lost differentiating a function

Click For Summary
SUMMARY

The discussion centers on differentiating a function in the context of non-linear optimization, specifically transitioning from equation 9.7.8 to 9.7.9 in the referenced PDF. The key equation discussed is g(λ) = f(xold + λp), where the derivative g'(λ) is derived as ∇f · p, using the Chain Rule. The confusion arises from understanding how the dot product is formed during differentiation. The clarification provided emphasizes the application of the Chain Rule and the representation of gradients and directional vectors.

PREREQUISITES
  • Understanding of non-linear optimization techniques
  • Familiarity with the Chain Rule in calculus
  • Knowledge of gradient vectors and dot products
  • Basic proficiency in mathematical notation and functions
NEXT STEPS
  • Study the Chain Rule in multivariable calculus
  • Learn about gradient descent methods in optimization
  • Explore the concept of directional derivatives
  • Review applications of dot products in vector calculus
USEFUL FOR

Students and professionals in mathematics, engineering, and computer science who are tackling non-linear optimization problems and require a solid understanding of differentiation techniques.

o_damhsa
Messages
2
Reaction score
0

Homework Statement

Homework Statement [/b]
Hello. I'm trying to solve a non-linear problem and I have been working through the notes on this pdf to try and understand the method before I use it but I get stuck at one of the steps. The pdf is here:
http://www.nrbook.com/a/bookfpdf/f9-7.pdf
I cannot follow how the author went from equation 9.7.8 to 9.7.9



Homework Equations


The author has the following equation:

[tex]g (\lambda) \equiv f(xold + \lambda p)[/tex]
He differentiates this function with respect to [tex]\lambda[/tex] and gets the following
[tex]g'(\lambda) = \nabla f \cdot p[/tex]


The Attempt at a Solution


My understanding is that the author differentiated with respect to [tex]\lambda[/tex] so they got the [tex]\nabla f[/tex] and then differentiated what was inside the function to get the [tex]p[/tex] value. But I don't see how it became a dot product, or am I just misreading it?
 
Physics news on Phys.org
Welcome to PF, o_damhsa


The pdf doesn't open for me, but here's how it looks from what you've said.

Call [tex]xold=\langle x_1,x_2,x_3\rangle[/tex] and [tex]p=\langle p_1,p_2,p_3 \rangle[/tex]

Then [tex]g(\lambda)=f(x_1+\lambda p_1,x_2+\lambda p_2,x_3+\lambda p_3)[/tex]

Now by the Chain Rule,

[tex]\frac{dg}{d\lambda}=<br /> \frac{\partial f}{\partial x}\frac{dx}{d\lambda}<br /> +\frac{\partial f}{\partial y}\frac{dy}{d\lambda}<br /> +\frac{\partial f}{\partial z}\frac{dz}{d\lambda}[/tex]

[tex]\frac{dg}{d\lambda}=<br /> \frac{\partial f}{\partial x} p_1<br /> +\frac{\partial f}{\partial y} p_2<br /> +\frac{\partial f}{\partial z} p_3[/tex]

[tex]\frac{dg}{d\lambda}=<br /> \left\langle \frac{\partial f}{\partial x},\frac{\partial f}{\partial y},\frac{\partial f}{\partial z}\right\rangle\cdot\langle p_1,p_2,p_3 \rangle[/tex]


[tex]\frac{dg}{d\lambda}=<br /> \nabla f\cdot p[/tex]
 
Hello Billy Bob,

Thank you very much for your reply! It makes perfect sense now :smile:

o_damhsa
 

Similar threads

Replies
6
Views
3K
  • · Replies 4 ·
Replies
4
Views
3K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 9 ·
Replies
9
Views
6K
Replies
1
Views
2K