Multiple-Choice Test

Newton’s Divided Difference Polynomial Method

Interpolation
COMPLETE SOLUTION SET

1. If a polynomial of degree n has more than n zeros, then the polynomial is

(A) oscillatory
(B) zero everywhere
(C) quadratic
(D) not defined

Solution
The correct answer is (B).

A unique polynomial of degree n or less passes through n  1 data points. Assume two
polynomials Pn  x  and Qn  x  go through n  1 data points,
 x0 , y 0  ,  x1 , y1  , ,  x n , y n 
Then
Rn  x   Pn  x   Qn  x 
Since Pn  x  and Qn  x  pass through all the n  1 data points,
Pn  xi   Qn  xi  , i  0,  , n
Hence
Rn  xi   Pn ( xi )  Qn  xi   0, i  0,  , n
The nth order polynomial Rn  x  has n  1 zeros. A polynomial of order n can have n  1 zeros
only if it is identical to a zero polynomial, that is,
Rn  x   0
Hence
Pn  x   Qn  x 
How can one show that if a second order polynomial has three zeros, then it is zero everywhere?
If R2  x   a 0  a1 x  a 2 x 2 , then if it has three zeros at x1 , x2 , and x3 , then
R2  x1   a0  a1 x1  a2 x12  0
R2  x2   a0  a1 x2  a2 x22  0
R2  x3   a0  a1 x3  a2 x32  0
Which in matrix form gives
1 x1 x12  a0  0
 
1 x 2 x 22   a1   0
1 x3 x32  a 2  0

The final solution a1  a 2  a3  0 exists if the coefficient matrix is invertible. The determinant of
the coefficient matrix can be found symbolically with the forward elimination steps of naïve
Gauss elimination to give
1 x1 x12 
 
det 1 x2 x22   x2 x32  x22 x3  x1 x32  x12 x3  x1 x22  x12 x2
1 x3 x32 
 
  x1  x2  x2  x3  x3  x1 
Since
x1  x 2  x3
the determinant is non-zero. Hence, the coefficient matrix is invertible. Therefore,
a1  a 2  a3  0 is the only solution, that is, R2  x   0.
 2. The following x - y data is given.
x 15 18 22
y 24 37 25
The Newton’s divided difference second order polynomial for the above data is given by
f 2 ( x)  b0  b1  x  15  b2  x  15 x  18
The value of b1 is most nearly
(A) –1.0480
(B) 0.14333
(C) 4.3333
(D) 24.000

Solution
The correct answer is (C).

Given
x0  15
x1  18
f 2  x0   24
f 2  x1   37
we have
f 2 ( x)  b0  b1  x  x0   b2  x  x0  x  x1 
Then
f 2 ( x0 )  b0  b1  x0  x0   b2  x0  x0  x0  x1 
 b0
f 2 ( x1 )  b0  b1  x1  x0   b2  x1  x0  x1  x1 
 b0  b1  x1  x0 
 f 2 ( x0 )  b1  x1  x0 
Thus,
f 2 ( x1 )  f 2 ( x0 )
b1 
x1  x0
37  24

18  15
 4.3333
 3. The polynomial that passes through the following x - y data

x 18 22 24
y ? 25 123

is given by
8.125 x 2  324.75 x  3237, 18  x  24
The corresponding polynomial using Newton’s divided difference polynomial is given by

f 2 ( x)  b0  b1  x  18  b2  x  18 x  22 
The value of b2 is

(A) 0.25000
(B) 8.1250
(C) 24.000
(D) not obtainable with the information given

Solution
The correct answer is (B).

Expanding,
f 2 ( x)  b0  b1  x  18  b2  x  18 x  22

 b0  b1 x  18b1  b2 x 2  40 x  396 
  b0  18b1  396b2    b1  40b2  x  b2 x 2
This needs to be the same as
8.125 x 2  324.75 x  3237
Hence
b2  8.125
 4. Velocity vs. time data for a body is approximated by a second order Newton’s divided
difference polynomial as

v t   b0  39.622 t  20  0.5540 t  20 t  15 , 10  t  20

The acceleration in m/s 2 at t  15 is

(A) 0.55400
(B) 39.622
(C) 36.852
(D) not obtainable with the given information

Solution
The correct answer is (C).

v t   b0  39.622 t  20  0.5540 t  20 t  15 , 10  t  20

d
a t    v t  
dt
 39.6221  0.5540 t  20   0.5540 t  15
 39.622  0.5540 t  20   0.5540 t  15 , 10  t  20
a 15  39.622  0.554015  20   0.554015  15
 39.622  2.7700
 36.852 m/s 2
5. The path that a robot is following on a x-y plane is found by interpolating four data points as

x 2 4.5 5.5 7
y 7.5 7.5 6 5
y  x   0.15238x 3  2.2571x 2  9.6048x  3.9000
The length of the path from x  2 to x  7 is
(A)  7.5  7.5 2   4.5  2 2   6  7.5 2   5.5  4.5
2
  5  6 2   7  5.5 2
7

(B) 
2
1  (0.15238x 3  2.2571x 2  9.6048x  3.9000) 2 dx
7

(C) 
2
1  (0.45714 x 2  4.5142 x  9.6048) 2 dx
7

 (0.15238x  2.2571x 2  9.6048 x  3.9000)dx

3
(D)
2

Solution
The correct answer is (C).

The length S of the curve y  x  from a to b is given by

b 2
 dy 
S   1   dx
a  dx 
where
a2
b7
giving
7 2
 dy 
S   1    dx
2  dx 
y  x   0.15238 x 3  2.2571x 2  9.6048 x  3.9000
dy
 0.45714 x 2  4.5142 x  9.6048
dx
Thus,
7

S   1  0.45714 x 2  4.5142 x  9.6048 dx  2

2
6. The following data of the velocity of a body is given as a function of time.
Time (s) 0 15 18 22 24
Velocity (m/s) 22 24 37 25 123
If you were going to use quadratic interpolation to find the value of the velocity at t  14.9
seconds, the three data points of time you would choose for interpolation are
(A) 0, 15, 18
(B) 15, 18, 22
(C) 0, 15, 22
(D) 0, 18, 24

Solution
The correct answer is (A).

We need to choose the three points closest to t  14.9 s that also bracket t  14.9 s . Although
the data points in choice (B) are closest to 14.9, they do not bracket it. This would be performing
extrapolation, not interpolation. Choices (C) and (D) both bracket t  14.9 s but they are not
the closest three data points.
Time (s) Velocity (m/s) How far is
t  14.9 s
0 22 14.9  0  14.9

15 24 14.9  15  0.1
18 37 14.9  18  3.1

22 25 14.9  22  7.1
24 123 14.9  24  9.1