This document contains 11 exercise questions related to computational errors and approximations from the Computational Methods in Engineering (ME6000) course at IIT Madras. The questions cover topics like converting decimal numbers to binary, evaluating Taylor series approximations of trigonometric functions, minimum and maximum integers representable on an 11-bit computer, condition numbers of functions, and analyzing sources of error like round-off errors in computations. The instructor's contact details are provided at the top.
This document contains 11 exercise questions related to computational errors and approximations from the Computational Methods in Engineering (ME6000) course at IIT Madras. The questions cover topics like converting decimal numbers to binary, evaluating Taylor series approximations of trigonometric functions, minimum and maximum integers representable on an 11-bit computer, condition numbers of functions, and analyzing sources of error like round-off errors in computations. The instructor's contact details are provided at the top.
This document contains 11 exercise questions related to computational errors and approximations from the Computational Methods in Engineering (ME6000) course at IIT Madras. The questions cover topics like converting decimal numbers to binary, evaluating Taylor series approximations of trigonometric functions, minimum and maximum integers representable on an 11-bit computer, condition numbers of functions, and analyzing sources of error like round-off errors in computations. The instructor's contact details are provided at the top.
Exercise Questions on Errors and Approximations Instructor: Ratna Kumar Annabattula email: ratna@iitm.ac.in August 12, 2014 1. Convert the following decimal numbers to their binary equivalents 23.45 0.5 267.32 2. Evaluate the true and approximate percentage relative error of estimating sin(x) at x = /3 through Taylor series around x = 0 using 2, 3 and 4 terms. 3. What are minimum and maximum numbers (both integers and oating point) that can be represented on a 11 bit computer. For oating point representation, choose your own break-up of bits for sign, mantissa and signed exponent. Also, discuss all the important aspects of representing real numbers on such a digital computer. 4. Show through an example that if a function is expanded using Taylor series about a point, then the function can be better approximated in its neighbourhood either by increasing the number of terms in the series or by narrowing the neighbourhood of interest. 5. If cos (2x) is expanded about x = 0 using Taylor series, the resulting series expansion will be continuous or discontinuous? 6. What is machine epsilon? Write an algorithm for calculating machine epsilon. 7. What is the range of mantissa for a normalized oating point representation scheme in a binary, decimal and octagonal system. 8. If you have a 4 bit computer that can only represent positive integers, what will be the result of 8 + 8 on this computer? 9. What is condition number of a function? Discuss about a function that is ill-conditioned at one point, but well-conditioned at another point. Choose an example different from the one discussed in the class. 10. Discuss the inuence of changing the number of bits in mantissa and exponent on the numbers repre- sented on a 11-bit computer discussed in Question. 3. 11. Show through an example that the round-off errors accumulate with increase in number of computations. 1