Journal of Physics: Conference Series

PAPER • OPEN ACCESS

Excel spreadsheet in teaching numerical methods

To cite this article: Harimi Djamila 2017 J. Phys.: Conf. Ser. 890 012093

View the article online for updates and enhancements.

This content was downloaded from IP address 179.61.164.55 on 21/09/2017 at 13:47

ICoAIMS 2017 IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1234567890
890 (2017) 012093 doi:10.1088/1742-6596/890/1/012093

Excel spreadsheet in teaching numerical methods

Harimi Djamila
Faculty of Engineering, Universiti Malaysia Sabah, Jalan UMS, Kota Kinabalu, Sabah
Malaysia
E-mail: [email protected]

Abstract. One of the important objectives in teaching numerical methods for undergraduates’
students is to bring into the comprehension of numerical methods algorithms. Although,
manual calculation is important in understanding the procedure, it is time consuming and prone
to error. This is specifically the case when considering the iteration procedure used in many
numerical methods. Currently, many commercial programs are useful in teaching numerical
methods such as Matlab, Maple, and Mathematica. These are usually not user-friendly by the
uninitiated. Excel spreadsheet offers an initial level of programming, which it can be used
either in or off campus. The students will not be distracted with writing codes. It must be
emphasized that general commercial software is required to be introduced later to more
elaborated questions. This article aims to report on a teaching numerical methods strategy for
undergraduates engineering programs. It is directed to students, lecturers and researchers in
engineering field.

1. Introduction
Numerical analysis is a subject of extreme importance to engineers in many applications. Numerical
methods are the study of set of procedure that uses numerical approximation for the solution of the
problems in mathematical analysis. This is by considering approximate error estimates. Numerical
methods became an independent mathematical discipline in the 20th Century. However, numerical
analysis originated in ancient times and slowly progressed and flourished when computers became
available in the late 1940s and early 1950s [1]. Currently numerical analysis is linked with computing
machines. At present, in various engineering courses, students are required to solve engineering
problems using numerical methods. This is important in acquiring useful engineering implications [2].
A course in numerical methods is also important in learning finite element method. In fact, a finite
element problem requires matrix using numerical methods. A mathematical software such Matlab
provides a platform for the students for fast learning by altering initial data without repeating all the
computations [2, 3].
Currently, many books on numerical methods using Matlab are available to the students such [4, 5,
6, 7] and many others. However, Matlab and other commercial software are costly. Further many
software are not user-friendly for uninitiated [8]. An attractive alternative for the students who are not
majoring in computer science is Excel spreadsheets. This alternative allows ready experimentation
with numerical algorithms [9]. In fact, this strategy is already used for learning finite element methods
[10]. Therefore, the objective of this article is to report on a teaching strategy, which could be
implemented for teaching numerical methods for undergraduates in Civil and Mechanical Engineering
degree programs. This article will illustrate the usefulness in using Excel spreadsheet for finding roots
using numerical methods. The spreadsheet approach helps the students to concentrate most of their

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
ICoAIMS 2017 IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1234567890
890 (2017) 012093 doi:10.1088/1742-6596/890/1/012093

efforts on thinking about the methods at hand, instead of turning their efforts to the need for
programming [10]. The developed strategy will not only benefit the student but also the lecturer. The
developed strategy is about learning by observation. Each time the student change for instance the
numerical problem, the student will acquire new knowledge by observations.
The developed strategy helps the lecturers to detect the evolution of the thinking process of their
students when applying numerical methods. The students also will certainly reach the conviction of
the importance of continuous learning in the future. Further, their Excel files will be useful for small
engineering applications or at least for checking the solution obtained via their own Matlab program
when possible or needed. The present illustration is based only on finding roots numerically; the
developed concept can be applied for other chapters with slight adjustment to the procedure used in
each chapter

2. Illustrative Example
The purpose of the selected example is to illustrate a new strategy in teaching numerical methods. This
strategy will be useful for both the lecturers and to the students. By using this strategy, the lecturers
will not only refer to the books for the preparation of their notes. They will be also learning about the
methods by observation while teaching their students. For the suggested strategy, the lecturers are
required to develop their own questions to be solved using numerical methods. These questions will be
solved using Excel spreadsheet. Such procedure will helps in questioning for instance the advantages
and limitation of the learned methods. Thus, the lecturer will be the author for his/her own tutorials.
Thus, the lecturer will be independently learning by observation. Further, this procedure will also help
the students to produce useful knowledge instead of consuming knowledge from books only. This is
important when the lecturers are at the beginning of their carriers specifically with engineering
backgrounds.
For illustration purpose, the chapter roots of equation were selected. The aim in teaching numerical
methods is not only learning numerical methods. It is also important that the students should be
involved in the thinking process when solving problems. For instance when using a bisection method,
the student can easily observe that the root can be found when the function cross the x-axis. Now the
lecturer may ask further question to the students how to write this information in an easy language,
which may be accepted by a computer. The computer has no eyes to observe the situation. The lecturer
can further inform the students that in many cases, when there is a root, the function cross the x-axis,
the function will change the sign from positive to negative value and vis versa. Thus if we randomly
select two initial values, xl located before the root and xu located after the root, in many cases f(xl).f(xu)
<0. Of course, there are exceptions, those are supposed to be given later to the student. The procedure
of thinking prior application of any numerical method is very important in grasping the concepts and
in understanding how the methods were developed. The strategy is for opening the students’ creativity
by observing first the simplicity of many numerical methods specifically at undergraduate level.
Probably, by following this strategy, a few students may suggest additional requirements from any
learned method in class or suggesting new methods.
Returning to our illustrative example, let us found the roots of a polynomial f(x)= -10x2 + 200x +
210. Any polynomial function of second order is recommended to be used firstly for illustration
purpose. This is to let most of students comfortable with the procedure. The students are expected also
to learn the difference between the exact and approximate solution. Additionally, they are supposed to
learn advantages and limitation of each method. The plot of the selected function is shown in Figure 1.
It is apparent from the figure that the present function has two real roots: (-1) and (21). It is important
to emphasise that the example was suggested without referring to any textbook and without an initial
knowledge of the expected results when using numerical methods. The suggested equation was only
checked to make sure that the function has two real roots to avoid exceptions. In the present
illustration, four basic numerical methods for finding roots were considered. These are Bisection
method, False-position method, Newton-Raphson method and Secant method. For the development of
the algorithm, the author referred to Chapra and Canale [11] Book. The first known drawback of

2
ICoAIMS 2017 IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1234567890
890 (2017) 012093 doi:10.1088/1742-6596/890/1/012093

numerical methods in finding roots is that, only one root can be estimated at a time within the selected
range. The solution will be mostly approximate. It is apparent from figure 1 that the two roots are
located within the interval [-6, 22].

3. Prediction of the first root of an equation using various numerical methods

For the estimation of the first root which is (-1) by using the bisection method, two initial guesses
[-6,4] were selected. It is apparent that the root is located within the selected interval. These two
values are initial guesses. Although, the bisection method is known to converge very slowly to the
solution, for the present case (figure 2), the method converged to the exact solution after one iteration
only. This situation occurred because the two selected initial guesses when using Bisection method
produced exact solution.

Figure 1. Selected function for finding roots.

The situation was different with the False-position method. It required 8 iterations to converge to
the exact solution. The results are illustrated in figure 3. The Excel file used for all the calculation will
be made available via ResearchGate. From the obtained results, it is apparent that the Bisection
method converged faster than the False-position method.

i Xl Xu Xr f(xl) f(xu) f(xr) f(xl)*f(xr) Approximate Error % Exact Error %

1 -6.0000 4.0000 -1.0000 -1350.0000 850.0000 0.0000 0.0000 / 0.00

Figure 2. Bisection method within the Interval [-6, 4].

i Xl Xu Xr f(xl) f(xu) f(xr) f(xl)*f(xr) Approximate Error % Exact Error %

1 -6.0000 4.0000 0.1364 -1350.0000 850.0000 237.0868 -320067.1488 / 113.64
2 -6.0000 0.1364 -0.7803 -1350.0000 237.0868 47.8478 -64594.5234 117.48 21.97
3 -6.0000 -0.7803 -0.9590 -1350.0000 47.8478 9.0067 -12159.0057 18.63 4.10
4 -6.0000 -0.9590 -0.9924 -1350.0000 9.0067 1.6730 -2258.5244 3.37 0.76
5 -6.0000 -0.9924 -0.9986 -1350.0000 1.6730 0.3100 -418.4810 0.62 0.14
6 -6.0000 -0.9986 -0.9997 -1350.0000 0.3100 0.0574 -77.5046 0.11 0.03
7 -6.0000 -0.9997 -1.0000 -1350.0000 0.0574 0.0106 -14.3530 0.02 0.00
8 -6.0000 -1.0000 -1.0000 -1350.0000 0.0106 0.0020 -2.6580 0.00 0.00
9 -6.0000 -1.0000 -1.0000 -1350.0000 0.0020 0.0004 -0.4922 0.00 0.00

Figure 3. False-Position method within the Interval [-6, 4].

For the next step, the derivative of the equation was determined. This was required for estimating
the root by using the Newton-Raphson method. The method requires only one initial guess. For
estimating the approximate root, the first selected initial guess was (-6). The procedure was also
repeated with an initial guess equal to (4). These two initial guesses were selected from the Bisection
method. This is without any consideration of other factors. Newton-Raphson method is a popular
method for finding the roots of an equation. It is widely known to converge fast to the solution.
However, it is an open method, so it may diverge in some cases; thus, no solution will be found.
Usually, when using the Newton-Raphson method, the initial guess is recommended to be close to the
root when possible. The obtained results from the present cases are illustrated in figure 4. It is apparent

3
ICoAIMS 2017 IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1234567890
890 (2017) 012093 doi:10.1088/1742-6596/890/1/012093

that the Newton-Raphson method converged faster than the False-position method. The situation is
different with the Bisection method.
For the next step, the Secant method was selected to find the root within the range of [-6, 4]. The
main difference between Secant and Newton method is this; the derivative of the function in Newton-
Raphson method is replaced by numerical differentiation equation. This is very useful when the
derivative of a function is difficult. Although, Secant method is an open method, it requires two initial
guesses. The method may not necessary converge to the solution. The obtained results are shown in
figure 5. From the figure, it is apparent that this method converged to the exact solution after 7
iteration.
i xi f(xi) f'(xi) Approximate Error % Exact Error % i xi f(xi) f"(xi) Approximate Error % Exact Error %
0 -6.0000 -1350.0000 320.0000 500.00 1 4.0000 850.0000 120.0000 500.00
2 -3.0833 -501.7361 261.6667 229.73 208.33
1 -1.7813 -177.9785 235.6250 236.84 78.13
3 -1.1659 -36.7666 223.3174 164.47 16.59
2 -1.0259 -5.7055 220.5181 73.63 2.59
4 -1.0012 -0.2711 220.0246 16.44 0.12
3 -1.0000 -0.0067 220.0006 2.59 0.00 5 -1.0000 0.0000 220.0000 0.12 0.00
4 -1.0000 0.0000 220.0000 0.00 0.00 6 -1.0000 0.0000 220.0000 0.00 0.00

Figure 4. Newton-Raphson method with an initial guess of (-6) and (4).

i xi-1 xi f(xi-1) f(xi) Approximate Error % Exact Error %
1 -6.0000 4.0000 -1350.0000 850.0000
2 4.0000 0.136364 850.0000 237.0868 2833.33 500.00
3 0.1364 -1.358166 237.0868 -80.0794 110.04 113.64
4 -1.3582 -0.980821 -80.0794 4.2156 38.47 -35.82
5 -0.9808 -0.999693 4.2156 0.0676 1.89 1.92
6 -0.9997 -1 0.0676 -0.0001 0.03 0.03
7 -1.0000 -1 -0.0001 0.0000 0.00 0.00

Figure 5. Secant method within the Interval [-6, 4].

4. Prediction of the second root of an equation using various numerical methods

The procedure of calculation is repeated for finding the second root of the equation, which is (21). The
Bisection method was firstly used. The selected initial guesses are [8, 22]. The obtained results are
shown in figure 6. This time, the method was only close to the solution. It required several iterations.
Given that, the approximate absolute percent relative error is only an approximate estimation of the
error, it is important to check that the function f(x) value is close to 0 at the final approximate root.
After completing the Bisection method, the procedure of calculation using the False-position method
was repeated with the selected interval [8,22]. The obtained results are shown in figure 7. Five
iterations were required to reach the solution. In this illustration, the False-position method reached the
exact solution. This was not the case of the Bisection method.

i Xl Xu Xr f(xl) f(xu) f(xr) f(xl)*f(xr) Approximate Error % Exact Error %

1 8.0000 22.0000 15.0000 1170.0000 -230.0000 960.0000 1123200.0000 / 28.57
2 15.0000 22.0000 18.5000 960.0000 -230.0000 487.5000 468000.0000 18.92 11.90
3 18.5000 22.0000 20.2500 487.5000 -230.0000 159.3750 77695.3125 8.64 3.57
4 20.2500 22.0000 21.1250 159.3750 -230.0000 -27.6563 -4407.7148 4.14 0.60
5 20.2500 21.1250 20.6875 159.3750 -27.6563 67.7734 10801.3916 2.11 1.49
6 20.6875 21.1250 20.9063 67.7734 -27.6563 20.5371 1391.8705 1.05 0.45
7 20.9063 21.1250 21.0156 20.5371 -27.6563 -3.4399 -70.6465 0.52 0.07
8 20.9063 21.0156 20.9609 20.5371 -3.4399 8.5785 176.1774 0.26 0.19
9 20.9609 21.0156 20.9883 8.5785 -3.4399 2.5768 22.1046 0.13 0.06
10 20.9883 21.0156 21.0020 2.5768 -3.4399 -0.4297 -1.1073 0.07 0.01
11 20.9883 21.0020 20.9951 2.5768 -0.4297 1.0740 2.7674 0.03 0.02
12 20.9951 21.0020 20.9985 1.0740 -0.4297 0.3222 0.3461 0.02 0.01
13 20.9985 21.0020 21.0002 0.3222 -0.4297 -0.0537 -0.0173 0.01 0.00
14 20.9985 21.0002 20.9994 0.3222 -0.0537 0.1343 0.0433 0.00 0.00

Figure 6. Bisection method within the Interval [8, 22].

4
ICoAIMS 2017 IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1234567890
890 (2017) 012093 doi:10.1088/1742-6596/890/1/012093

i Xl Xu Xr f(xl) f(xu) f(xr) f(xl)*f(xr) Approximate Error % Exact Error %

1 8.0000 22.0000 19.7000 1170.0000 -230.0000 269.1000 314847.0000 / 6.19
2 19.7000 22.0000 20.9401 269.1000 -230.0000 13.1438 3537.0057 5.92 0.29
3 20.9401 22.0000 20.9974 13.1438 -230.0000 0.5745 7.5506 0.27 0.01
4 20.9974 22.0000 20.9999 0.5745 -230.0000 0.0250 0.0144 0.01 0.00
5 20.9999 22.0000 21.0000 0.0250 -230.0000 0.0011 0.0000 0.00 0.00

Figure 7. False-position method within the Interval [8, 22].

The procedure for finding the root (21) was also made with the Newton-Raphson method. For the
first calculation, the selected initial guess was (8). The second selected initial guess was (22). The
results of both cases are plotted in figure 8. When the value of the initial guess was (8), the method
failed to find the expected root which (21). However, the first root was found. This case is widely
known when using Newton-Raphson method. Interestingly, the approximate error converged to (0)
after 8th iteration. When an initial guess of (22) was selected, the Newton-Raphson method required
little iteration to converge to the exact expected solution.

i xi f(xi) f"(xi) Approximate Error % Exact Error % i xi f(xi) f"(xi) Approximate Error % Exact Error %
1 8.0000 1170.0000 40.0000 61.90 1 22.0000 -230.0000 -240.0000 2300.00

2 -21.2500 -8555.6250 625.0000 137.65 201.19 2 21.0417 -9.1840 -220.8333 4.55 2204.17
3 21.0001 -0.0173 -220.0016 0.20 2200.01
3 -7.5610 -1873.8872 351.2200 181.05 136.00
4 21.0000 0.0000 -220.0000 0.00 2200.00
4 -2.2256 -284.6613 244.5127 239.72 110.60
5 21.0000 0.0000 -220.0000 0.00 2200.00
5 -1.0614 -13.5536 221.2287 109.68 105.05 6 21.0000 0.0000 -220.0000 0.00 2200.00
6 -1.0002 -0.0375 220.0034 6.13 104.76 7 21.0000 0.0000 -220.0000 0.00 2200.00
7 -1.0000 0.0000 220.0000 0.02 104.76 8 21.0000 0.0000 -220.0000 0.00 2200.00
8 -1.0000 0.0000 220.0000 0.00 104.76 9 21.0000 0.0000 -220.0000 0.00 2200.00

Figure 8. Newton-Raphson method with an initial guess of (8) and (22).

The selected initial guess in this case was close to the root. The approximate root was (21). It is worth
making comparison between False-position and Newton-Raphson methods by considering the
approximate error. It is apparent that the Newton-Raphson method with an initial guess of (22)
reached faster the root than the False-position method. So, what will be the case with the Secant
method? The procedure was repeated to determine the root by using two initial guesses. Those are (8)
and (22). The results are plotted in figure 9. It is apparent that the method converge to the exact
solution after 6 iteration. It is obvious that the initial guess (22) is closer to the solution than (8). The
convergence was faster than the Newton-Raphson method with an initial guess equal to (8) but slower
than the Newton-Raphson method with an initial guess equal to (22).
i xi-1 xi f(xi-1) f(xi) Approximate Error % Exact Error %
1 8.0000 22.0000 1170.0000 -230.0000
2 22.0000 19.7 -230.0000 269.1000 11.68 4.76
3 19.7000 20.94009 269.1000 13.1438 5.92 6.19
4 20.9401 21.00377 13.1438 -0.8303 0.30 0.29
5 21.0038 20.99999 -0.8303 0.0023 0.02 0.02
6 21.0000 21 0.0023 0.0000 0.00 0.00
Figure 9. Secant method within the Interval [8, 22].

5. Further numerical analysis

In the last section, two initial guesses were selected to bracket both roots. This is by using Bisection
and False position methods. The selected range was [-2, 22]. The results are plotted in figures 10 and
11. Although, the Bisection method converged slowly to the first root, the situation was different with
False-position method. For this last case, it was expected that the method will converge to either (-1)
or (21). However, this was not the case. The f(x) values of the two selected initial guesses were the
same. Therefore, it was not possible to find out the approximate solution given there is a division by
zero.

5
ICoAIMS 2017 IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1234567890
890 (2017) 012093 doi:10.1088/1742-6596/890/1/012093

i Xl Xu Xr f(xl) f(xu) f(xr) f(xl)*f(xr) Approximate Error %

1 -2.0000 22.0000 10.0000 -230.0000 -230.0000 1210.0000 -278300.0000 /
2 -2.0000 10.0000 4.0000 -230.0000 1210.0000 850.0000 -195500.0000 150.00
3 -2.0000 4.0000 1.0000 -230.0000 850.0000 400.0000 -92000.0000 300.00
4 -2.0000 1.0000 -0.5000 -230.0000 400.0000 107.5000 -24725.0000 -300.00
5 -2.0000 -0.5000 -1.2500 -230.0000 107.5000 -55.6250 12793.7500 -60.00
6 -1.2500 -0.5000 -0.8750 -55.6250 107.5000 27.3438 -1520.9961 -42.86
7 -1.2500 -0.8750 -1.0625 -55.6250 27.3438 -13.7891 767.0166 -17.65
8 -1.0625 -0.8750 -0.9688 -13.7891 27.3438 6.8652 -94.6651 -9.68
9 -1.0625 -0.9688 -1.0156 -13.7891 6.8652 -3.4399 47.4336 -4.62
10 -1.0156 -0.9688 -0.9922 -3.4399 6.8652 1.7181 -5.9103 -2.36
11 -1.0156 -0.9922 -1.0039 -3.4399 1.7181 -0.8595 2.9567 -1.17
12 -1.0039 -0.9922 -0.9980 -0.8595 1.7181 0.4296 -0.3693 -0.59
13 -1.0039 -0.9980 -1.0010 -0.8595 0.4296 -0.2149 0.1847 -0.29
14 -1.0010 -0.9980 -0.9995 -0.2149 0.4296 0.1074 -0.0231 -0.15
15 -1.0010 -0.9995 -1.0002 -0.2149 0.1074 -0.0537 0.0115 -0.07

Figure 10. Bisection method with initial guesses [-2, 22].

i Xl Xu Xr f(xl) f(xu) f(xr) f(xl)*f(xr)

1 -2.0000 22.0000 #DIV/0! -230.0000 -230.0000 #DIV/0! #DIV/0!
2 #DIV/0! #DIV/0! #DIV/0! #DIV/0! #DIV/0! #DIV/0! #DIV/0!
3 #DIV/0! #DIV/0! #DIV/0! #DIV/0! #DIV/0! #DIV/0! #DIV/0!
Figure 11. False-position method with initial guesses [-2, 22].

For the case of Newton-Raphson, the method converged to the root which was closer to the initial
guess, which is (-1). The results are plotted in figure 12. For the Secant method, similar problem
occurred as for False-position method. There is a division by zero. This is illustrated also in figure 12.
Therefore, further calculation was not possible. It is important to mention such cases cannot be
necessary learned from book. The benefit in using Excel is clearly shown in the present illustrative
example which is learning by observation.

i xi f(xi) f"(xi) Approximate Error % i xi-1 xi f(xi-1) f(xi) Approximate Error %

1 -2.0000 -230.0000 240.0000 1 -2.0000 22.0000 -230.0000 -230.0000
2 -1.0417 -9.1840 220.8333 92.00 2 22.0000 #DIV/0! -230.0000 #DIV/0! #DIV/0!
3 -1.0001 -0.0173 220.0016 4.16 3 #DIV/0! #DIV/0! #DIV/0! #DIV/0! #DIV/0!
4 -1.0000 0.0000 220.0000 0.01 4 #DIV/0! #DIV/0! #DIV/0! #DIV/0! #DIV/0!
5 -1.0000 0.0000 220.0000 0.00 5 #DIV/0! #DIV/0! #DIV/0! #DIV/0! #DIV/0!
(a) Newton-Raphson method (b) Secant method
Figure 12. Newton-Raphson method with an initial guess (-2)
& Secant method with initial guesses [-2, 22].

At this stage, students are ready for more advanced numerical methods in finding roots and for writing
their first Matlab codes. For any unexpected issue with their Matlab program, the student can always
refer back to Excel file to check carefully the solution.

6. Conclusion
This article focuses on using Excel spreadsheet in numerical methods as a teaching strategy for
undergraduate students. The strategy is about learning many methods using one single equation to
observe the advantages and limitations of each numerical method. The developed strategy helps the
lecturers to detect the evolution of the thinking process of their students when applying numerical
methods. An illustrative example about finding roots using four numerical methods was suggested.
The developed concept can be applied for other chapters with slight adjustment to the procedure used
in each chapter. The developed strategy goes beyond learning from textbooks only. The students will
learn from their own experience with the first application used in numerical methods. The developed

6
ICoAIMS 2017 IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1234567890
890 (2017) 012093 doi:10.1088/1742-6596/890/1/012093

strategy will be of great help to the new lecturers by acquiring new knowledge with each application.
The students also will certainly reach the conviction of the importance of long life learning. Further,
their Excel files will be useful for small engineering applications or at least for checking the solution
obtained via their own Matlab program when needed.

References
[1] Gutknecht M H 2010 Seminar on Numerical analysis in Zurich – 50 years ago (Seminar für
Angewandte Mathematik, Eidgenössische Technische Hochschule, Rämistr)
[2] Kim A S, Park C, Park SH 2003 Development of Web-based engineering numerical software
(WENS) Using MATLAB Applications to Linear Algebra (Computer Applications in
Engineering Education) pp 67-74
[3] Martín-Caraballo A M & Tenorio-Villalón A F 2015. Teaching Numerical Methods for Non-
linear Equations with GeoGebra-Based Activities ( Mathematics Education vol 2) pp 53-65
[4] Matthews JH. and Fink KD 2004 Numerical Methods using MATLAB.
(New Jersey: Pearson.) 4th Edition.
[5] Young T & Martin J M 2017 Introduction to Numerical Methods and Matlab Programming for
Engineers (Department of Mathematics: Ohio University Athens)
[6] Yang WY, Cao W, Chung T, Morris J 2005 Applied Numerical Methods Using MATLAB (John
Wiley & Sons, Inc).
[7] Chapra SC 2008 Applied Numerical Methods with MATLAB for Engineers and Scientists. (New
York: McGraw Hill) 2nd Edition
[8] Zuraini O, Abdul S b S 2010 Proc. Conf. on Numerical Method Calculator for Engineering
Education (NUMCeE Malaysian Technical Universities Conference on Engineering and
Technology Bayview Hotel, Melaka, Malaysia).
[9] Wyk C J V 2006 Using Spreadsheets to Learn Numerical Methods. (eJSiE vol 1) pp148-157.
[10] Kand T and L Morgan 2005 Proc. On The Application of EXCEL in Teaching Finite Element
Analysis to Final Year Engineering Students (ASEE/A aeE 4th Global Colloquium on
Engineering Education by Australasian Association for Engineering Education).
[11] Chapra SC and Canale R P 2015 Numerical Methods for Engineers (New York: McGraw Hill)
7th Edition.