Program To Do Numerical Analysis in MATLAB Language Assignment Solution.

function [flag, approx] = progb820474(f,a,b,tol)

% [flag, b,r]= proga820474(x)

% Author: Amanda Lauen

% Date: October 1, 2021

% Description:

%

% Input

% f = an annonymous function supplied by the user

% a = represents the x0 term respectfully, given by user

% b = represents the x1 term, respectfully, given by user

% tol = stoppong tolerance where the algorithm stops when the

% function is less than the tolerance

%

% Outputs

% flag = flag to signify program executed properly or not

% flag = 0 --> program runs without error

% flag = 1 --> input is out of range

% flag = 2 --> the algorithm fails to converge

% approx = approximation calculated using the Secant Method

% Sample

% [flag, approx]= progb820474();

% [flag, approx]= progb820474();

% Check that the bounds of the problem are correct

if b <= a

    flag = 1;

    approx = -99;

    return;

end

% Secant Method Code:

% f = @(x) function

% x = 1.1:0.1:5; --> Range

% y = f(x);

% plot(x,y);

% x0 = ____;

% x1 = x0+f(x0);

% x2 = x1-f(x1*(x1-x0)/(f(x1-f(x0))

% x0 = x1;

% x1=x2; --> Update

% Repeat this step over and over to see if it is converging or not

% x2 = x1-f(x1*(x1-x0)/(f(x1-f(x0))

% x0 = x1;

% x1=x2; --> Update

% MAKE A MAX ITERATION (Choose 101) --> Check to see if it converges

% To make it, count each iteration within a loop (for loop to make

% sure it doesn't become an infinite loop)

% If the algorithm diverges, then flag = 2;

% Store the different variables to distinguish them; determine if

% they are next to each other; if x1-x0 bounces around or oscillates

% between two values, then it diverges (flag 2)

% Convergence/Divergence test on functions (Review from prev math

% classes)

% If x0 less than or equal to x1 and the summation of x1 is

% convergence, then x0 is convergent (if greater than, they diverge)

%

% Notes for Overleaf

% How I validated that my program giving accurate calculation

% Check to see if f(x) = 0 exists

% Check to see if the function is a feasable solution

% Program should only return the output variables and should not output or

% print anything else

% Notes from Riley

% Make judgement calls.

% Don't have to check if letters are entered or not

% All numbers are coming in, no tests

% All functions are coming in, no tests

% If a and b are bad guesses, it diverges/ don't worry about bad inputs

% for a and b

% How to detect it converges or not

% Use while loop

% Stop criteria: abs(b-a)/abs(b) < tol;

% Not Converging: if you go 1000 iterations and it's greater than

% tol, it diverges

% What is a good cap for this? --> Have to defend it in the report

% Counting approach is better than going off to infinity

% a and b have to be close enough to the root

% What to do if root is close or equal to zero: Have to discuss that in

% the report. Have a fail safe to stop the program.

% Convergence: Superlinear or slightly off of linear

% Note: The questions on existence of a root and checking of a solution

% os feasible are discssion items for the report and they are not PROGRAM

% SPECIFICATIONS: DO NOT CODE THIS. REPORT ON THIS ONLY!!!!!!

% Where to get the root: answer, how do you know it is feasible?

%% We define a maximum number of iterations, initial error and initial flag

max_iters = 500; % 500 iterations

error = Inf; % initial error

flag = 0; % runs without errors

approx =f(a);

%% Begin with Secant Method Algorithm

n = 1; % iterations counter

while error >= tol % execute algorithm while the error is higher than the tolerance

    x0 = a;

    x1 = b;

    x2 = x1 - (x1-x0)*f(x1) / (f(x1)-f(x0)); % calculate x2

    aold = a;

    bold = b;

    % Update

    a = x1;

    b = x2;

    error = abs(b-a)/abs(a);

 %% Enable the following line if you want to check the error per iteration

% fprintf("Iter = %d, aold = %.4f, bold = %.4f, x2 = %.4f, anew = %.4f, bnew = %.4f, error = %.8f\n", n, aold, bold, x2, a, b, error);

% fprintf("Iteration %d, error: %.8f\n", n, error);

    % Increment step

    n = n + 1;

    approx = b;

    % If the number if iterations reaches the max. number of iterations

    % allowed, it means that the problem did not find a solution

    if n >= max_iters

        flag = 2;

        break;

    end

end