1. Not finding help here? Sign up for a free 30min tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Someone help me

  1. Feb 12, 2006 #1

    F.B

    User Avatar

    I need help with these sort of questions where you have given a set of
    points and you must determine the equation, linear ones are easy but ones
    that are quadratic or cubic. So could go about and tell me how to determine
    equations of polynomials. Heres the set of points i am given.

    x.....y
    -3....-110
    -2....-21
    -1.....4
    0.....7
    1.....6
    2....-5
    3....-56
    The dots separate the columns because they are supposed to be in a table of values.

    So can anyone help me determine an equation for this. I know i have to find the differences in the y-column, so the fourth difference of the y-column are constant at -24, but what do i do now. I know the equation is in the form y=ax^4 + bx^3 + cx^2 + dx + e, but what do i do with the 24 and how do i solve the rest.
     
  2. jcsd
  3. Feb 12, 2006 #2

    arildno

    User Avatar
    Science Advisor
    Homework Helper
    Gold Member
    Dearly Missed

    you are to find an interpolating polynomial, right?

    Remember that you can always interpolate with a polynomial of degree one less than the number of points you've got.
    that is:
    You may use a sixth-degree polynomial here, if it is actually possible to interpolate with a polynomial of less degree, that will become apparent when solving the equations for the sixth-degree polynomial
     
  4. Feb 12, 2006 #3

    Fermat

    User Avatar
    Homework Helper

    You don't bother with the 24 now. You just used that to determine the order of the polynomial.

    You have a 4th order polynomial with 5 unknowns, a,b,c,d,e.

    Use 5 points from your data set, 5 (x,y)-values, to create 5 simultaneous eqns in a,b,c,d,e.
    Then use matrix methods to solve for the unknowns, i.e. a,b,c,d,e.
     
  5. Feb 12, 2006 #4

    HallsofIvy

    User Avatar
    Staff Emeritus
    Science Advisor

    Since you talk about the "fourth difference" apparently you know about "Newton's divided difference formula". It is essentially the same as Taylors series but works with finite differences:
    [tex]P(x)= y(x_0)+ \Delta_1(x- x_0)+ \frac{\Delta_2}{2}(x-x_0)(x-x_1)+ ...+ \frac{\Delta_n}{n!}(x- x_0)(x-x_1)...(x- x_n)[/tex]
    x0, x1,etc are the x values starting at some point (in your case, x0= -3, x1= -2,etc.) The terms [itex]\Delta_1[/itex], etc are the divided differences (since your x-values are all unit steps apart, that's just the differences). Notice that you do not have powers of (x-x0) but products of (x-x0[/sub](x- x1) etc.

    The computation involved is a bit easier than solving 4 equations for 4 unknowns. Also this has the advantage that it is easier to program.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?