Está en la página 1de 4

HOMEWORK ASSIGNMENT 3 EGM 3344

Numerical Errors & root finding

Due: Monday, Sept. 26, 2011----2:59 pm


Problems from Chapra book chapter 4 For all of these problems, feel free to use a simple Matlab program to calculate function values (like using a calculator) when the indicated approach is by hand. Problem Approach Comments 4.2 Hand Hint: First number is an integer, second number is a floating point. Write out solution equation by hand. Do calculations in Matlab if desired. 4.4 Matlab Name your Matlab program macheps.m 4.5 Matlab Skip the challenge question. 4.8 Hand Only do 3-digit arithmetic part. Write out all equations by hand and use Matlab or calculator for intermediate results. Chopping means truncate to only 3 digits after each intermediate calculation step. 4.13 Hand 4.16 Matlab Evaluate the remainder term in the Taylor series at x = 2 to estimate Et 4.22 S1 4.2 Matlab hand Study the code on pp.116-117 of 3rd ed or pp.105-106 of 2nd ed.. See problem statement below.

Convert the following base-2 numbers to base 10: a) 1011001; b) 0.01011; c) 110.01001

4.4 For computers the machine epsilon can also be thought as the smallest number that when added to one gives a number greater than l. An algorithm based on this idea can be developed as Step 1: Set =1. Step 2: If is less than or equal to 1, then go to Step 5. Step 3: = /2. Step 4: Return to Step 2. Step 5: = 2 x . Write your own M-file based on this algorithm to determine the machine epsilon .Validate the result by comparing it with the value computed with the built-in function eps. 4.5 In a fashion similar to Prob.4.4, develop your own M-file to determine the smallest positive real number used in Matlab. Base your algorithm on the notion that your computer will

2 be unable to reliably distinguish between zero and a quantity that is smaller than this number. Note that the result you obtain will differ from the value computed with realmin. Challenge question: Investigate the results by your code and those obtained with realmin. 4.8 The derivative of f(x)=1/(1-3x2) is given by 6x/(1-3x2)2 Do you expect to have difficulties evaluating this function at x= 0.577? Try it using 3- and 4digit arithmetic with chopping. 4.13 Use zero- through third-order Taylor series expansions to predict f(3) for f(x) =25x3 6x2 + 7x - 88 using a base point at x=l. Compute the true percent relative error t for each approximation. Discuss the meaning of the results. 4.16 Use forward and backward difference approximations of O(h) and a centered difference approximation of O(h2) to estimate the first derivative of the function examined in Prob.4.13. Evaluate the derivative at x=2 using a step size of h= 0.25. Compare your results with the true value of the derivative. Interpret your results on the basis of the remainder term of the Taylor series expansion. 4.22 Repeat Example 4.5 but for f(x)= cos(x) at x = /6. (note: Example 4.5 in the 2nd Edition is the same as in the 3rd Edition)

S1. (Nested form) Evaluate f(x) = x3 - 7.5x2 + 11.2x + 2.8 = ((x - 7.5)x + l l.2)x + 2.8 at x = 3.94 using 3 significant (3S) arithmetic and rounding, in both of the given forms. The latter, called the nested form, is usually preferable since it minimizes the number of operations and thus the effect of rounding. (Note: for computing 7.5x2, you get exact x2=15.5236 first. Rounding it to 3S, you get 15.5. Thus 7.5x2 is 7.5*15.5 whose exact value is 116.25. Rounding it to 3S, you should get 116. Rounding should be applied after each operation.) Answers 4.2 a) 4.4 89; b) 0.34375; c) 6.28125

>> macheps ans = 2.2204e-016 >> eps ans = 2.2204e-016

4.5
4.9407e-324

3 (or any extremely small value on the order of e-320 solution may vary slightly from one computer to the next) 4.8 True value: f(0.577) = 2,352,911 3-digit result with chopping: f(0.577) = 216,250 => chopping to 216,000 t = 90.8% 4.13 True value: f(3) = 554 Zero order: f(3) -62, t = 111.19% First order: f(3) 78, t = 85.92% Second order: f(3) 354, t = 36.10% Third order: f(3) 554, t = 0% 4.16 True value: f(2) = 283 Forward: f(2) = 320.5625, Et = 37.5625 Backward: f(2) = 248.5625, Et = 34.4375 Centrered: f(2) = 284.5625, Et = 1.5625 Et based on remainder term in Taylor series: Forward: Et = 36 Backward: Et = 36 Centrered: Et = 1.56 4.22

S1: Unnested f =-8.10 Nested: f =-8.2