Math 171 - Alexiades
                            Lab 11
      Finding roots - Newton-Raphson Method
A. Newton-Raphson root-finder
  Write a code Newton.m that implements the Newton-Raphson method for finding a root of F(x) = 0.
  It requires F(x) to be differentiable with computable derivative DF(x).
  The user should provide: x0: initial guess, maxIT: max number of iterations, and TOL: tolerance for convergence.
  Newton converges fast typically, so maxIT=20 usually suffices.
  For convenience, you may set maxIT=20 and TOL = 1.e-14 in the code so only x0 needs to be provided as argument.
 
  • The code should:   accept input   x0 ,
          print out the iterates:   n   xn   Fn   (neatly, in columns, use format: %d %f %e ),
      and, upon convergence, print out the value of the root, the residual (%e format), and how many iterations it took.
  • The (formula for) F(x) and DF(x) should be coded in a subprogram:   function [Fn , DFn] = FCN( xn )

    B. To do:
      1. Debug the code on a simple problem with known solution, e.g. F(x) = x2 − 9.
      2. Use your Newton code to find all three roots of the cubic: F(x) = x3 + x2 − 3x − 3
        Try various x0, like: 2, 1, 0.5, 0.1, 0.0, −0.5, −1, etc.
        Do you notice anything unexpected/interesting? what?
      3. Also try some big x0:  100, 1000, −100, etc
        Do you notice anything unexpected/interesting? what?
      4. Try some other maxIT and TOL.
      5. Answer the questions:
        Q1: What are the three roots you found ? How many iterations did each one take ? with what x0 ?
        Q2: Is Newton more efficient than Bisection? how did you decide?
        Q3: What is the effect of changing the initial guess x0 ?
        Q4: What is the effect of changing maxIT ?
        Q5: What is the effect of changing the tolerance ?
      6. Now find the real root of "Newton's cubic": x3 − 2x − 5 = 0
        (comment out previous formulas in FCN and insert new ones).
        This is the only equation Newton ever bothered to solve with his method, in 1671 !

    C. Prepare Lab11 for submission. In a plain text file Lab11.txt , insert:
      1. Name, Date, Lab11
       %=====================================%
      2. Answers to the questions in B.5  (each labeled and separated by %------------------------%)
       %=====================================%
      3. What is the (real) root of "Newton's cubic" ? in how many iterations, with what x0 ?
       %=====================================%
      4. Insert your Newton.m code.
       %=====================================%
      5. Insert your FCN.m code.

    Using a PLAIN TEXT editor, edit the file Lab11.txt to clean it up,
    leaving only the correct version of what's asked for in this Lab.

    Then submit it on Canvas.