Computer Engineering 3SK3

Numerical Analysis

Academic year 2019-2020, term 2


Instructor: Xiaolin Wu, ITB-A315

Extension: 24190



Lectures: 3 hours/week

Tutorial:  1 hour/week       




Chu, Xiaoxuan,; office hours: Monday, 1~4pm.  

ITB A204 Desk16.


Liang, Runchen,; office hours: Tuesday 1~4 pm.

ITB A204, desk 5


Qian, Jingjing,; office hours: Thursday 1~4 pm.

 ITB A203.


Luo, Fangzhou,; office hours: Wednesday 1~4 pm.

ITB A103


The course outline:  3sk3\CompEng_3SK3_outline_Winter2020.pdf


Lecture Notes: part1.pdf; part2.pdf; part3.pdf; part4.pdf; LinearProg.pdf; part5.pdf; Interpolation; part6.pdf; part7.pdf; 3sk3\interpolation.pdf


Tutorials: Tutorial_1 3SK3 2013.pdf, 3sk3\Tutorial_2.pdf, 3sk3\Tutorial_3.pdf                         


Textbook:  “Numerical Methods for Engineers, Sixth edition”, by S. C. Chapra and R. P. Canale, McGraw Hill, 2010.


Course Objectives:


 Main topics:

·       Roots of Equations

·       Linear Algebraic Equations

·       Optimization Techniques

·       Data Fitting

·       Numerical Differentiation and Integration

·       Differential Equations


Format: The course consists of class lecture sessions, tutorial session and a laboratory component. The lab component of the course consists of two small projects that students can do on their own time schedule.


Assessment scheme:


Tests: 20%

Final: 60%

Projects: 20%



Course Project (TBA)




OLD COURSE MATERIALS (for references only):


Sample midterm exam Midterm11.pdf


Answers to midterms 1 and 2: Answers


Sample questions for the final exam. Sample exam questions




Marks of Assignments 1 and 2 A1A2 Marks.pdf


Assignment 1 (Due January 17, 2014):


Prove the following two definitions of machine precision are equivalent.


1.     Machine precision is the maximum relative error.

2.     Machine precision is the smallest number e such that fl(1+e)>1.


What is the machine precision for IEEE 64-bit floating point format?



Assignment 2 (Due January 31, 2014)


Prove function  has a unique root in interval [0,1].  If the bisection search method

is used to find the root, how many iterations are required to ensure that the error is below ?

How many iterations does the Newton-Raphson method require to achieve the same precision?

How many iterations does the False-point method require to achieve the same precision?




Sample questions: 3sk3\Sample Q1.pdf,  3sk3\Sample Q2.pdf, 3sk3\Sample Q3.pdf, 3sk3\Sample Q4.pdf,


More exercise questions are in


Textbook:  “Numerical Methods for Engineers, Sixth edition”, by S. C. Chapra and R. P. Canale, McGraw Hill, 2010.



Q13.2 (ed.5) (answer x = 0.91692)

Q13.3 (ed.5) (the 3rd estimate x* = 0.9443)

Q13.5 (ed.5) (the result after three iterations is x*=1.047716)

Q13.11 (ed.5) (x = -0.5867)

Q14.2 (ed.5) ()

Q14.3 (ed.5)

answer for (b) is




Q14.6 (ed.5) (one iteration suffices to converge)


Yes more sample questions:


17.4 ed.5 (17.3 ed.6), 17.9 ed.5 (17.8 ed.6), 17.13 ed.5 (17.12 ed.6), 17.14 ed.5, 17.16 ed.5 (17.18 ed.6)


18.5 ed.5 (18.6 ed.6), 18.7 ed.5 (18.7 ed.6)


21.1, 21.2 (both ed.5 and ed.6)


22.1, 22.3 (both ed.5 and ed.6)


23.1, 23.3, 23.4 (both ed.5 and ed.6)


25.2 ed.5