Sturm-Liouville Theory - Partial Differential Equations | Lecture 23

แชร์
ฝัง
  • เผยแพร่เมื่อ 4 ธ.ค. 2024

ความคิดเห็น • 6

  • @JesusSportsNature
    @JesusSportsNature หลายเดือนก่อน +1

    Thank you for these videos!

  • @sschmachtel8963
    @sschmachtel8963 3 หลายเดือนก่อน

    Yeah actually, the higher the eigenvalues the similar the eigenfunctions are to sine and cosine especially for bigger x. I guess thats for many but not all depending on p and q

    • @jasonbramburger
      @jasonbramburger  3 หลายเดือนก่อน +1

      What do you mean by similar to sine and cosine? How is this measured? Many famous eigenfunctions are polynomial, so not similar at all since sine and cosine are transcendental functions.

    • @sschmachtel8963
      @sschmachtel8963 3 หลายเดือนก่อน

      @@jasonbramburger eigenvalues are similar for some of them. My experience. Has something to do with the number of zeros on the interval and the further to the right boundary the more eigenfunction resembles sine or Cousine. I just think why this is .... maybe only when q term is Zero? What I Studied was Diffusion covection with quadraticly in space dependence.... dy2dx2 -3x2 dydx + lambda y =0

    • @jasonbramburger
      @jasonbramburger  3 หลายเดือนก่อน

      @@sschmachtel8963 Here is an example of eigenvalues and eigenfunctions that are very different from sine and cosine examples: www.global-sci.org/oldweb/jms/freedownload/v50n2/pdf/502-101.pdf

    • @sschmachtel8963
      @sschmachtel8963 3 หลายเดือนก่อน

      @@jasonbramburger thx. Yeah I just remembered when I was deriving some analytical solutions for convection diffusive systems and how rich all the SL theory and also generally orthogonal polynomials, including general fourier series is. And that eigenvalue similarity was probably only for some specific systems, anyway using those eigenvalues for lambda100 or more gave a much better reconstruction of the initial condition.
      Really interesting is also fitting of transients, where you have the sum of exponentials. Didnt yet quite understand why DMD seems to not like those, but works better for complex eigenvalues or periodic in time functions. That I namely also tried to do:: Fit eigenvalues out of measured data, but that failed horribly, I remember reading it somewhere that this is very difficult because lamdas can interswitch, and I dont exactly know if constraints in the form lamda1