‫‪Numerical Analysis – Spring 2021 Final Exam‬‬

‫‪Important‬‬
‫• الـ ‪ PDF‬ده مجهود طالب مش دكتور‪ ،‬وارد يكون فيه أخطاء‪ .‬فـ ياريت محدش يعتمد اي حاجة هنا كـ‬
‫‪model answer‬‬
‫• السؤال اللي قدامه "☹" معناه إني معرفتش أحله أو إني مش واثق من إلمعلومة اللي عندي خالص‬
‫• أي تعديل في الـ ‪ PDF‬هينزل على الدرايف ‪[Click Here] ‬‬
Numerical Analysis – Spring 2021 Final Exam Page |2

Question 1) [10 Marks]

a) Solve the following system of equation using LU Decomposition with partial pivoting: [5 marks]

𝟐𝒙𝟏 + 𝒙𝟐 + 𝒙𝟑 = 𝟓
𝟒𝒙𝟏 − 𝟔𝒙𝟐 = −𝟐
−𝟐𝒙𝟏 + 𝟕𝒙𝟐 + 𝟐𝒙𝟑 = 𝟗
Step 1: Augmented Matrix A & B Step 2: Swap R1 <-> R2 (because 4 > 2)
A: B: P21A:
|2 1 1 | |5| |4 -6 0 |
|4 -6 0 | |-2| |2 1 1 | m21 = 2/4
| -2 7 2 | |9| | -2 7 2 | m31 = -2/4

Step 3: Apply R2 = R2 - R1(m21) Step 4: Apply R3 = R3 - R1(m31)

P21A: P21A:
|4 -6 0 | |4 -6 0 |
|0 4 1 | |0 4 1 |
| -2 7 2 | |0 4 2 | m32 = 4/4

Step 5: Apply R3 = R3 - R2(m32) Step 6: P21B → B

P21A: |5| |-2|
|4 -6 0 | |-2| |5|
|0 4 1 | |9| |9|
|0 0 1 |

Step 7: Get L Matrix Step 8: Apply LC = B:

L: |1 0 0 | | C1 | | -2| c1 = -2
|1 0 0 | | 0.5 1 0 | | C2 | = | 5 | c2 = 6
| 0.5 1 0 | | -0.5 1 1 | | C3 | | 9 | c3 = 2
| -0.5 1 1 |

Step 9: Apply UX = C:
|4 -6 0 | | X1 | | -2| c1 = 1
|0 4 1 | | X2 | = | 6 | c2 = 1
|0 0 1 | | X3 | | 2 | c3 = 2

- Ahmed Huzain
Numerical Analysis – Spring 2021 Final Exam Page |3

b) Use Gauss Jordan with partial pivoting to solve the system of equations given in (a) [5 marks]

𝟐𝒙𝟏 + 𝒙𝟐 + 𝒙𝟑 = 𝟓
𝟒𝒙𝟏 − 𝟔𝒙𝟐 = −𝟐
−𝟐𝒙𝟏 + 𝟕𝒙𝟐 + 𝟐𝒙𝟑 = 𝟗
Step 1: Augmented Matrix Step 2: Swap R1 <-> R2 (because 4 > 2)
A: A:
|2 1 1 5 | |4 -6 0 -2 |
|4 -6 0 -2 | |2 1 1 5 |
| -2 7 2 9 | | -2 7 2 9 |

Step 3: Divide R1 by 4 Step 4: R2 = R2 + -2 * R1

|1 -1.5 0 -0.5 | |1 -1.5 0 -0.5 |
|2 1 1 5 | |0 4 1 6 |
| -2 7 2 9 | | -2 7 2 9 |

Step 5: R3 = R3 + 2 * R1 Step 6: Divide R2 by 4

|1 -1.5 0 -0.5 | |1 -1.5 0 -0.5 |
|0 4 1 6 | |0 1 0.3 1.5 |
|0 4 2 8 | |0 4 2 8 |

Step 7: R1 = R1 + 1.5 * R2 Step 8: R3 = R3 + -4 * R2

|1 0 0.4 1.8 | |1 0 0.4 1.8 |
|0 1 0.3 1.5 | |0 1 0.3 1.5 |
|0 4 2 8 | |0 0 1 2 |

Step 9: R1 = R1 + -0.3 * R3 Step 10: R2 = R2 + -0.3 * R3

|1 0 0 1 | |1 0 0 1 | x1 = 1
|0 1 0.3 1.5 | |0 1 0 1 | x2 = 1
|0 0 1 2 | |0 0 1 2 | x3 = 2

- Ahmed Huzain
Numerical Analysis – Spring 2021 Final Exam Page |4

Question 2) [12 Marks]

a) Use the bisection method (halving) to find the smallest positive root of the following equation
f(x) = −𝟐𝒙𝟔 − 𝟏. 𝟔𝒙𝟒 + 𝟏𝟐𝒙 − 𝟏 until the estimated error Ea= 5%. Employ the initial guess of x1 = 0 and xu=l
[ 5 marks ]

i Xl F(xl) xu F(xu) xr F(xr) ea%

1 0 -1 1 7.4 0.5 4.869 ----
2 0 -1 0.5 4.869 0.25 1.993 100%
3 0 -1 0.25 1.993 0.125 0.5 100%
4 0 -1 0.125 0.5 0.063 -0.244 98.413%
5 0.063 -0.244 0.125 0.5 0.094 0.128 32.979%
6 0.063 -0.244 0.094 0.128 0.079 -0.052 18.987%
7 0.079 -0.052 0.094 0.128 0.087 0.044 9.195%
8 0.079 -0.052 0.087 0.044 0.083 -0.004 4.819%

Root = 0.083

b) Write C++ program for getting the root for problem 2a using recursive function

#include <iostream>
#include <cmath>
using namespace std;
const double eps = 5.0;
int iter = 0;
double xr = 0, xrOld = 0, error = 0;
double f(double x){
return -2 * pow(x, 6) - 1.6 * pow(x, 4) + 12 * x - 1;
}
double bisect(double xl, double xu, double error, int iter, double xr){
xrOld = xr;
xr = (xl + xu) / 2;
error = abs((xr - xrOld) / xr) * 100;
if (f(xl) * f(xr) > 0) xl = xr;
else if (f(xl) * f(xr) < 0) xu = xr;
else return xr;
iter++;
if(error > eps) bisect(xl, xu, error, iter, xr);
else return xr;
}
int main(){
float xl = 0, xu = 1;
cout << "Root =" << bisect(xl, xu, error, iter, xr) << endl;
return 0;
}

- Ahmed Huzain
Numerical Analysis – Spring 2021 Final Exam Page |5

c) Find the minimum value of 𝒇(𝒙, 𝒚) = (𝒙 − 𝟐)𝟐 + (𝒚 − 𝟑)𝟐 starting at x = 1 and y = 1, using the conjugate
gradiant method [5 marks]

Step 1: Get df/dx & df/dy

df/dx: 2 * (x - 2) = -2
df/dy: 2 * (y - 3) = -4

𝑑𝑓 𝑑𝑓 −4
Step 2: Calculate gradiant F = 𝑑𝑥 𝑖 + 𝑑𝑦 𝑗 = 𝑡𝑎𝑛−1 (−2) = 63.435

Step 3: H matrix:
𝑑2𝑓 𝑑2𝑓
| |
𝑑𝑥 2 𝑑𝑥𝑑𝑦
𝑑2𝑓 𝑑2𝑓
| |
𝑑𝑦𝑑𝑥 𝑑𝑦 2

Step 4:Substitute in H matrix:

| 2 0 |
| 0 2 |

Step 5: Calculate Det(H)

| H | = 2 (2) - 0 (0) = 4

Step 6: Final result:

Det|H| > 0 AND d2f/dx^2 > 0 → local min

- Ahmed Huzain
Numerical Analysis – Spring 2021 Final Exam Page |6

Question 3) [13 Marks]

a) Determine the real root of f(x) = −𝒙𝟑 + 𝟕. 𝟖𝟗𝒙 + 𝟏𝟏. Using the Secant Method Use an initial guess of xi-1
= 3 and x0 = 5 and iterate until ea% < 0.1. [5 marks]

i Xi-1 F(xi-1) xi F(xi) ea%

1 3 7.67 5 -74.55 ----
2 5 -74.55 3.187 3.775 56.8%
3 3.187 3.775 3.274 1.738 2.65%
4 3.274 1.738 3.348 -0.112 2.21%
5 3.348 -0.112 3.344 -0.01 0.14%
6 3.344 -0.01 3.344 -0.01 0.02%

Root = 3.344

b) Write pseudocode for calculating roots of the previous function in (3a). [3 marks]

FUNCTION Secant(a, b, imax, es)

iter = 0
DO
x = a – (f(a)*(b-a)/f(b)-f(a))
iter = iter + 1
ea = abs((xr - xrold)/xr)*100
b = a
a = x
IF ea < es OR iter ≥ imax EXIT
END DO
Secant = x
END Secant

- Ahmed Huzain
Numerical Analysis – Spring 2021 Final Exam Page |7

c) Write the algorithm for Gauss elimination using pivoting to solve system of equation and then write C++
code for it. [5 marks]

Algorithm

1- Forward elimination
| 𝑎11 𝑎12 𝑎13 𝑏1 | If 𝑎11 < 𝑎21 𝑂𝑅 𝑎11 < 𝑎31 𝑇𝐻𝐸𝑁 𝑠𝑤𝑎𝑝
| 𝑎21 𝑎22 𝑎23 𝑏2 |
| 𝑎31 𝑎32 𝑎33 𝑏3 |

| 𝑎11 𝑎12 𝑎13 𝑏1 |

| 0 𝑎22 𝑎23 𝑏2 | If 𝑎22 < 𝑎32 𝑇𝐻𝐸𝑁 𝑠𝑤𝑎𝑝
| 0 𝑎32 𝑎33 𝑏3 |

| 𝑎11 𝑎12 𝑎13 𝑏1 |

| 0 𝑎22 𝑎23 𝑏2 |
| 0 0 𝑎33 𝑏3 |

2- Backward substitution
𝑏3
𝑥3 =
𝑎33
𝑏3 − 𝑎23 ∗ 𝑥3
𝑥2 =
𝑎22
𝑏1 − 𝑎12 ∗ 𝑥2 − 𝑎13 ∗ 𝑥3
𝑥1 =
𝑎11

Code:
main():
int main(){
float _a[3][4];
float _m21 = 0, _m31 = 0, _m32 = 0;
cout << "Enter matrix" << endl;
for (int i = 0; i < 3; i++)
{
for (int j = 0; j < 4; j++)
{
cout << "row" << i << j << ": ";
cin >> _a[i][j];
}
}
GJE(_a, _m21, _m31, _m32);
}

- Ahmed Huzain
Numerical Analysis – Spring 2021 Final Exam Page |8

Swap():
void swap(float _a[][4], int r1, int r2){
cout << "Swap R" << r1 << " <-> R" << r2 << endl;
for(int i = 0; i < 4; i++){
float temp = _a[r1][i];
_a[r1][i] = _a[r2][i];
_a[r2][i] = temp;
}
}

GJE():
void GJE(float _a[][4], float &m21, float &m31, float &m32){
// Check if pivot is larger than the elements below it
if(abs(_a[0][0]) < abs(_a[1][0]) || abs(_a[0][0]) < abs(_a[2][0]))
if(abs(_a[1][0]) > abs(_a[2][0]))
swap(_a, 0, 1);
else swap(_a, 0, 2);

m21 = _a[1][0]/_a[0][0]; // Calculate m21

m31 = _a[2][0]/_a[0][0]; // Calculate m31

for(int j = 0; j < 4; j++){

_a[1][j] -= _a[0][j] * m21; // R2 = R2 - R1(m21)
_a[2][j] -= _a[0][j] * m31; // R3 = R3 - R1(m31)
}

// Check if m22 is larger than the element below it

if(abs(_a[1][1]) < abs(_a[2][1]))
swap(_a, 1, 2);

m32 = _a[2][1]/_a[1][1]; // Calculate m32

for(int j = 0; j < 4; j++)

_a[2][j] -= _a[1][j] * m32; // R3 = R3 - R2(m32)

float x3 = _a[2][3] / _a[2][2]; // Get x3

float x2 = (_a[1][3] - (_a[1][2] * x3)) / _a[1][1]; // Get x2
float x1 = (_a[0][3] - ((_a[0][1] * x2) + (_a[0][2] * x3))) / _a[0][0]; // Get x1
cout << "Gauss Jordan Result" << endl << "X1 = " << x1 << endl << "X2 = "
<< x2 << endl << "X3 = " << x3 << endl;
}

- Ahmed Huzain
Numerical Analysis – Spring 2021 Final Exam Page |9

Question 4) [15 Marks]

a) Use the golden-section search to find the maximum of
f (x) = −𝟏. 𝟓𝒙𝟔 − 𝟐𝒙𝟒 + 𝟏𝟐𝒙 Within the interval x1 = 0 and Xu= 2. [5 marks]

5 ‫ انا افترضت‬,iterations ‫السؤال هنا مش محدد عدد‬

i Xl F(xl) x2 F(x2) x1 F(x1) xu F(xu) d

1 0 0 0.764 8.188 1.236 4.816 2 -104 1.236
2 0 0 0.472 5.548 0.764 8.188 1.236 4.816 0.764
3 0.472 5.548 0.764 8.188 0.944 8.678 1.236 4.816 0.472
4 0.764 8.188 0.944 8.678 1.056 8.105 1.236 4.816 0.292
5 0.764 8.188 0.876 8.656 0.944 8.678 1.056 8.105 0.18

b) Use Newton method to find the real root to four decimals of the following polynomials
f(x) = 𝒙𝟒 + 𝒙𝟑 − 𝟒𝒙𝟐 − 𝟑𝒙 + 𝟑. Use an initial guess of x0 = 5 and iterate until ea < 10%. [4 marks]

i Xi F(xi) F`(xi) Ea%

1 5 638 532 ----
2 3.801 197.455 229.596 31.544%
3 2.941 59.831 101.173 29.242%
4 2.35 17.336 46.679 25.149%
5 1.979 4.486 23.92 18.747%
6 1.791 0.83 15.275 10.497%
7 1.737 0.064 13.119 3.109%

Root = 1.933

- Ahmed Huzain
Numerical Analysis – Spring 2021 Final Exam P a g e | 10

c) Write a C++ program for getting the root using Simple fixed point for a function f(x) = -x3 + 7.89x + 11. Use
initial guess ofx0 = 1 and iterate until ea < 2 % [3 marks]

#include <iostream>
#include <cmath>
using namespace std;
double eps = 2;
int iter = 0;
double xiPlus1 = 0, xi = 0, error = 0;

double f(double x){

return cbrt(7.89*x+11);
}

double fixedpoint(double x){

xi = x;
xiPlus1 = f(xi);
cout << "i=" << iter << " | Xi= " << x << " | Xi+1= " << xiPlus1 << " | E= %" <<
error << endl;
error = abs((xiPlus1 - xi) / xiPlus1) * 100;
iter++;
if (error > eps) fixedpoint(xiPlus1);
return xi;
}

int main(){
double xo = 1;
double root;
root = fixedpoint(xo);
cout << "root= " << root << endl;
return 0;
}

𝒂 𝟏 𝟏 𝒂
d) If [ ]=4,[ ] = 12. Use Cramer’s Rule to find values of x1, x2 [3 marks]
𝒃 𝟕 𝟑 𝒃
𝒙𝟏 + 𝒙𝟐 = 𝒂
𝟑𝒙𝟏 + 𝟕𝒙𝟐 = 𝒃

1 1
[ ]
3 7
D=1*7–1*3=4

D1 = 4, D2 = 12
𝐷1 4
X1 = =4 =1
𝐷
𝐷2 12
X2 = = =3
𝐷 4

- Ahmed Huzain