Problem 4.4: Gauss Elimination Method Problem 4.9: Gauss–Jordan Elimination Method 1 0 0.8 −1.1 5 ← −(0.8)[0 0 1 1.75 −2.

em 4.9: Gauss–Jordan Elimination Method 1 0 0.8 −1.1 5 ← −(0.8)[0 0 1 1.75 −2.5] 8.1.2 Two-point backward difference formula for first derivative
0 1 −0.4 1.8 −1 ← −(−0.4)[0 0 1 1.75 −2.5]
Solve the following system of equations using the Gauss elimination method. Solve the following system of equations using the Gauss–Jordan elimination ൦ ൪ = f ᇱᇱ (x ୧ ) ଶ f ᇱᇱᇱ (x୧ ) ଷ f (ସ) (x ୧ ) ସ
0 0 1 1.75 −2.5 ← f(x୧ିଵ ) = f(x୧ ) − f ᇱ (x୧ )h + h − h + h +⋯
method. 2! 3! 4!
Given system of equations: 0 0 8.2 −10.4 29 ← −(8.2)[0 0 1 1.75 −2.5]
f ᇱᇱ (ξ) ଶ
2x₁ + x₂ − x₃ + 2x₄ = 0 Given system of equations: 1 0 0 −2.5 7 f(x୧ିଵ ) = f(x୧ ) − f ᇱ (x୧ )h + h
x₁ − 2x₂ + x₃ − 4x₄ = 3 4x₁ + 3x₂ + 2x₃ + x₄ = 17 0 1 0 2.5 −2 2!
൦ ൪ f(x୧ ) − f(x୧ିଵ ) f ᇱᇱ (ξ)
3x₁ − x₂ − 2x₃ − x₄ = −3 2x₁ − x₂ + 2x₃ − 4x₄ = 11 0 0 1 1.75 −2.5 → f ᇱ (x୧ ) = + h
−x₁ + 2x₂ + x₃ − 2x₄ = 13 x₁ + 2x₂ − 2x₃ − x₄ = 8 0 0 0 −24.75 49.5 h 2!
f ᇱᇱ (ξ)
−2x₁ + 4x₂ + 5x₃ − x₄ = 15 Normalize row 4 by dividing by -24.75 and eliminate above: truncation error = h = O(h)
The system of equations in matrix form is: 2!
First, form the augmented matrix: f(x୧ ) − f(x୧ିଵ )
2 1 −1 2 xଵ 0 1 0 0 −2.5 7 ← −(−2.5)[0 0 0 1 −2] → f ᇱ (x୧ ) = + O(h)
1 −2 1 −4 xଶ 3 0 1 0 2.5 −2 ← −(2.5)[0 0 0 1 −2] h
൦ ൪൦ ൪ = ൦ ൪ 4 3 2 1 17 ൦ ൪ =
3 −1 −2 −1 xଷ −3 2 −1 2 −4 11 0 0 1 1.75 −2.5 ← −(1.75)[0 0 0 1 −2] 8.1.3 Two-point central difference formula for first derivative
൦ ൪
−1 2 1 −2 xସ 13 1 2 −2 −1 8 0 0 0 1 −2 ←
ୟ ଵ ୟ ଷ ୟ ିଵ
1 0 0 0 2 f ᇱᇱ (x୧ ) ଶ f ᇱᇱᇱ (ξଵ ) ଷ
With mଶଵ = ୟమభ = ଶ = 0.5, mଷଵ = ୟయభ = ଶ = 1.5, mସଵ = ୟరభ = ଶ = −0.5, the −2 4 5 −1 15 f(x୧ାଵ ) = f(x୧ ) + f ᇱ (x୧ )h + h + h
భభ భభ భభ 0 1 0 0 3 2! 3!
first pass with Gaussian elimition yields: The pivot element aଵଵ = 4. Normalize row 1 by dividing by 4. ൦
0 0 1 0 1
൪ f ᇱᇱ (x୧ ) ଶ f ᇱᇱᇱ (ξଶ ) ଷ
f(x୧ିଵ ) = f(x୧ ) − f ᇱ (x୧ )h + h − h
2 1 −1 2 xଵ 0 1 0.75 0.5 0.25 4.25 ← 0 0 0 1 −2 2! 3!
2 −1 2 −4 11 ← −(2)[1 0.75 0.5 0.25 4.25] f ᇱᇱᇱ (ξଵ ) ଷ f ᇱᇱᇱ (ξଶ ) ଷ
0 −2.5 1.5 −5 xଶ 3 ൦ ൪ = Final solution: x₁ = 2, x₂ = 3, x₃ = 1, x₄ = −2 f(x୧ାଵ ) − f(x୧ିଵ ) = 2f ᇱ (x ୧ )h + h + h
൦ ൪൦ ൪ = ൦ ൪ 1 2 −2 −1 8 ← −(1)[1 0.75 0.5 0.25 4.25] 3! 3!
0 −2.5 −0.5 −4 xଷ −3
−2 4 5 −1 15 ← −(−2)[1 0.75 0.5 0.25 4.25] f(x୧ାଵ ) − f(x୧ିଵ ) f ᇱᇱᇱ (ξଵ ) f ᇱᇱᇱ (ξଶ ) ଶ
0 2.5 0.5 −1 xସ 13 Chapter 8: ᇱ (x )
→f ୧ = −ቆ + ቇh
ᇲ ᇲ
1 0.75 0.5 0.25 4.25 2h 2.3! 2.3!
ୟయమ ିଶ.ହ ୟరమ ଶ.ହ 8.1 Finite Difference Formulas of First Derivative
Next, with mଷଶ = = = 1, mସଶ = = = −1, the second pass with 0 −2.5 1 −4.5 2.5 f(x୧ାଵ ) − f(x୧ିଵ )
ୟᇲమమ ିଶ.ହ ୟᇲమమ ିଶ.ହ ൦ ൪ → f ᇱ (x୧ ) = + O(hଶ )
0 1,25 −2.5 −1.25 3.75 8.1.1 Two-point forward difference formula for first derivative 2h
Gaussian elimination yields:
0 5.5 6 −0.5 23.5 f ᇱᇱᇱ (ξଵ ) f ᇱᇱᇱ (ξଶ ) ଶ
2 1 −1 2 xଵ 0 f ᇱᇱ (x ୧ ) ଶ f ᇱᇱᇱ (x୧ ) ଷ f (ସ) (x ୧ ) ସ truncation error = − ቆ + ቇ h = O(hଶ )
Normalize row 2 by dividing by -2.5 and eliminate above/below: f(x୧ାଵ ) = f(x୧ ) + f ᇱ (x୧ )h + h + h + h +⋯ 2.3! 2.3!
0 −2.5 1.5 −5 xଶ 3 2! 3! 4!
൦ ൪൦ ൪ = ൦ ൪
0 0 −2 1 xଷ −6 1 0.75 0.5 0.25 4.25 ← −(0.75)[0 1 −0.4 1.8 −1] where h = x୧ାଵ − x୧ 8.1.4 Three-point forward and backward difference formulas for first derivative
0 0 2 −6 xସ 16 0 1 −0.4 1.8 −1 ←
൦ ൪ = Note: above is also Taylor series formula of f(x୧ାଵ ). Forward:
ᇲᇲ
ୟరయ ଶ 0 1,25 −2.5 −1.25 3.75 ← −(1.25)[0 1 −0.4 1.8 −1]
Next, with mସଷ = ୟᇲᇲ
= = −1, the third pass with Gaussian elimination 0 5.5 6 −0.5 23.5 ← −(5.5)[0 1 −0.4 1.8 −1] f ᇱᇱ (x୧ ) ଶ f ᇱᇱᇱ (ξଵ ) ଷ
యయ ିଶ Rewritten as:
yields: 1 0 0.8 −1.1 5 f(x୧ାଵ ) = f(x୧ ) + f ᇱ (x୧ )h + h + h
f ᇱᇱ (ξ) ଶ 2! 3!
0 1 −0.4 1.8 −1 f(x୧ାଵ ) = f(x୧ ) + f ᇱ (x୧ )h + h f ᇱᇱ (x୧ ) f ᇱᇱᇱ (ξଶ )
2 1 −1 2 xଵ 0 ൦
0 0 −2 −3.5 5
൪ 2! f(x୧ାଶ ) = f(x୧ ) + f ᇱ (x୧ )2h + (2h)ଶ + (2h)ଷ
0 −2.5 1.5 −5 xଶ 3 where ξ is a value of x between x୧ and x୧ାଵ 2! 3!
൦ ൪ ൦x ൪ = ൦ ൪ 0 0 8.2 −10.4 29 ᇱᇱ ᇱ
f ᇱᇱᇱ (ξଵ ) ଷ f ᇱᇱᇱ (ξଶ )
0 0 −2 1 ଷ −6 f(x୧ାଵ ) − f(x ୧ ) f (ξ) 4f(x୧ାଵ ) − f(x ୧ାଶ ) = 3f(x୧ ) + 2f (x୧ )h + 4 h − (2h)ଷ
0 0 0 −5 xସ 10 Normalize row 3 by dividing by -2 and eliminate above/below: → f ᇱ (x୧ ) = − h 3! 3!
h 2! 4f(x୧ାଵ ) − f(x ୧ାଶ ) = 3f(x୧ ) + 2f ᇱ (x୧ )h − 2hO(hଶ )
ଵ଴ ି଺ିଵ(ିଶ) f ᇱᇱ (ξ) −3f(x୧ ) + 4f(x୧ାଵ ) − f(x୧ାଶ )
Using back substitution, xସ = ିହ = −2, x ଷ = = 2, xଶ = truncation error = − h = O(h) f ᇱ (x୧ ) = + O(hଶ )
ଷି(ିହ)(ିଶ)ି(ଵ.ହ)(ଶ)
ିଶ
଴ି(ଶ)(ିଶ)ି(ିଵ)(ଶ)ି(ଵ)(ସ)
2! 2h
= 4, xଵ = = 1. Thus, the solution is f(x୧ାଵ ) − f(x ୧ )
ିଶ.ହ ଶ → f ᇱ (x୧ ) = + O(h) Similar, three point backward
[xଵ xଶ xଷ xସ ] = [1 4 2 −2] h

3f(x୧ ) − 4f(x୧ିଵ ) + f(x୧ିଶ ) f(x୧ାଵ ) − f(x୧ ) ௕

2-point O(h) Third derivative
f ᇱ (x୧ ) = + O(hଶ ) f ᇱ (x୧ ) = ‫ = )݂(ܫ‬න ݂(‫ݔ݀)ݔ‬
2h forward h Method Formula Truncatio ௔
Example 8-3: f(x) = x ଷ . Calculate the first derivative at point x = 3 numerically 3-point −3f(x୧ ) + 4f(x୧ାଵ ) − f(x୧ାଶ ) O(hଶ ) n Error = ݂(‫ݔ‬ଵ )(‫ݔ‬ଶ − ‫ݔ‬ଵ ) + ݂(‫ݔ‬ଶ )(‫ݔ‬ଷ − ‫ݔ‬ଶ ) + ⋯ + ݂(‫ݔ‬ே )(‫ݔ‬ேାଵ − ‫ݔ‬ே )
with the 3-point forward difference formula, using: f ᇱ (x୧ ) = ே
forward 2h 4-point ᇱᇱᇱ (x
−f(x୧ ) + 3f(x୧ାଵ ) − 3f(x୧ାଶ ) + f(x ୧ାଷ ) O(h) = ෍[݂(‫ݔ‬௜ )(‫ݔ‬௜ାଵ − ‫ݔ‬௜ )]
a. points x = 3, x = 4, x = 5 2-point f(x୧ ) − f(x୧ିଵ ) O(h) forward
f ୧) =
hଷ ௜ୀଵ
f ᇱ (x୧ ) =
x 3 4 5 backwar h When the subintervals have the same width h:
5-point f ᇱᇱᇱ (x୧ ) O(hଶ )
f′(x) 27 64 125 d
forward −5f(x୧ ) + 18f(x୧ାଵ ) − 24f(x୧ାଶ ) + 14f(x୧ାଷ ) ே
௕
−3f(3) + 4f(4) − f(5) −3.27 + 4.64 − 125 3-point 3f(x୧ ) − 4f(x୧ିଵ ) + f(x୧ିଶ ) O(hଶ ) −3f(x୧ାସ ) ‫ = )݂(ܫ‬න ݂(‫ = ݔ݀)ݔ‬ℎ ෍ ݂(‫ݔ‬௜ )
f ᇱ (3) = = = 25 f ᇱ (x୧ ) = =
2.1 2 backwar 2h 2hଷ ௔ ௜ୀଵ
25 − 27 d 4-point −f(x୧ିଷ ) + 3f(x୧ିଶ ) − 3f(x୧ିଵ ) + f(x୧ ) O(h)
error = ฬ ฬ . 100% = 7.41% f ᇱᇱᇱ (x୧ ) = 9.2.1 Midpoint
27 2-point f(x ୧ାଵ ) − f(x୧ିଵ ) O(hଶ ) backwar hଷ
8.2 Finite Difference Formulas of Second Derivative f ᇱ (x୧ ) = d
௕ ௕
ܽ+ܾ ܽ+ܾ
central 2h ‫ = )݂(ܫ‬න ݂(‫ = ݔ݀)ݔ‬න ݂ ൬ ൰ ݀‫ ݂ = ݔ‬൬ ൰ (ܾ − ܽ)
௔ ௔ 2 2
8.2.1 Three-point central difference formula for the second derivative 4-point f(x୧ିଶ ) − 8f(x ୧ିଵ ) + 8f(x୧ାଵ ) − f(x୧ାଶ ) O(hସ ) 5-point f ᇱᇱᇱ (x୧ ) O(hଶ )
f ᇱ (x ୧ ) = backwar 3f(x ୧ିସ ) − 14f(x୧ିଷ ) + 24f(x୧ିଶ ) − 18f(x୧ିଵ ) 9.2.2 Composite Midpoint
central 12h
f ᇱᇱ (x ୧ ) ଶ f ᇱᇱᇱ (x୧ ) ଷ f (ସ) (x ୧ ) ସ d +5f(x୧ )
f(x୧ାଵ ) = f(x୧ ) + f ᇱ (x୧ )h + h + h + h +⋯ = Domain [ܽ, ܾ] is divided into N subintervals by ܰ + 1 points ‫ݔ‬ଵ , ‫ݔ‬ଶ , … , ‫ݔ‬ேାଵ
2! 3! 4! Second derivative 2hଷ
f ᇱᇱ (x ୧ ) ଶ f ᇱᇱᇱ (x୧ ) ଷ f (ସ) (x ୧ ) ସ where ‫ݔ‬ଵ = ܽ, ‫ݔ‬ேାଵ = ܾ
f(x୧ିଵ ) = f(x୧ ) − f ᇱ (x୧ )h + h − h + h +⋯ Method Formula Truncatio 4-point ᇱᇱᇱ
−f(x୧ିଶ ) + 2f(x ୧ିଵ ) − 2f(x୧ାଵ ) + f(x୧ାଶ ) O(hଶ ) ௕
2! 3! 4! n Error f (x୧ ) =
f ᇱᇱ (x୧ ) ଶ f (ସ) (x୧ ) ସ central 2hଷ ‫ = )݂(ܫ‬න ݂(‫ݔ݀)ݔ‬
f(x୧ାଵ ) + f(x୧ିଵ ) = 2f(x୧ ) + 2 h +2 h +⋯ f(x୧ ) − 2f(x ୧ାଵ ) + f(x୧ାଶ ) ௔
2! 4! 3-point ᇱᇱ (x O(h) 6-point f(x୧ିଷ ) − 8f(x୧ିଶ ) + 13f(x୧ିଵ ) − 13f(x୧ାଵ ) O(hସ ) ‫ݔ‬ଵ + ‫ݔ‬ଶ ‫ݔ‬ଶ + ‫ݔ‬ଷ
f(x୧ିଵ ) − 2f(x୧ ) + f(x୧ାଵ ) forward
f ୧) = = ݂൬ ൰ (‫ݔ‬ଶ − ‫ݔ‬ଵ ) + ݂ ൬ ൰ (‫ݔ‬ଷ − ‫ݔ‬ଶ ) + ⋯
f ᇱᇱ (x୧ ) = + O(hଶ ) hଶ central +8f(x୧ାଶ ) − f(x ୧ାଷ ) 2 2
hଶ f ᇱᇱᇱ (x୧ ) = ‫ݔ‬ே + ‫ݔ‬ேାଵ
4-point 2f(x୧ ) − 5f(x୧ାଵ ) + 4f(x୧ାଶ ) − f(x ୧ାଷ ) O(hଶ ) 8hଷ
8.2.2 Three-point forward and backward difference formula for the second +݂൬ ൰ (‫ݔ‬ேାଵ − ‫ݔ‬ே )
f ᇱᇱ (x୧ ) = 2
derivative forward hଶ ே
‫ݔ‬௜ + ‫ݔ‬௜ାଵ
Forward: 3-point f(x୧ିଶ ) − 2f(x୧ିଵ ) + f(x୧ ) O(h) Chapter 9: Integration = ෍[݂( )(‫ݔ‬௜ାଵ − ‫ݔ‬௜ )]
f ᇱᇱ (x୧ ) = 2
backwar hଶ ௜ୀଵ
f(x୧ ) − 2f(x ୧ାଵ ) + f(x୧ାଶ ) 9.1.1 Rectangle method
f ᇱᇱ (x୧ ) = + O(hଶ ) d When the subintervals have the same width h:
hଶ ௕ ௕
Backward: 4-point 2f(x୧ ) − 5f(x୧ିଵ ) + 4f(x୧ିଶ ) − f(x ୧ିଷ ) O(hଶ ) ‫ = )݂(ܫ‬න ݂(‫ = ݔ݀)ݔ‬න ݂(ܽ)݀‫ ܾ()ܽ(݂ = ݔ‬− ܽ) ௕ ே
f ᇱᇱ (x୧ ) = ‫ݔ‬௜ + ‫ݔ‬௜ାଵ
backwar hଶ ௔ ௔ ‫ = )݂(ܫ‬න ݂(‫ = ݔ݀)ݔ‬ℎ ෍ ݂( )
f(x୧ିଶ ) − 2f(x୧ିଵ ) + f(x୧ ) ௕ 2
f ᇱᇱ (x
୧) = + O(hଶ ) d ‫ = )݂(ܫ ݎ݋‬න ݂(ܾ)݀‫ ܾ()ܾ(݂ = ݔ‬− ܽ)
௔ ௜ୀଵ
hଶ ௔
3-point f(x୧ିଵ ) − 2f(x୧ ) + f(x୧ାଵ ) O(hଶ ) 9.3.1 Trapezoidal
f ᇱᇱ (x୧ ) = 9.1.2 Composite rectangle
central hଶ ௕ ௕
First derivative ݂(ܾ) − ݂(ܽ)
Domain [ܽ, ܾ] is divided into N subintervals by ܰ + 1 points ‫ݔ‬ଵ , ‫ݔ‬ଶ , … , ‫ݔ‬ேାଵ ‫ = )݂(ܫ‬න ݂(‫ = ݔ݀)ݔ‬න (݂(ܽ) + (‫ ݔ‬− ܽ) )݀‫ݔ‬
5-point f ᇱᇱ (x
୧) O(h ) ସ
௔ ௔ ܾ−ܽ
Method Formula Truncatio −f(x୧ିଶ ) + 16f(x୧ିଵ ) − 30f(x୧ ) + 16f(x୧ାଵ ) − f(x୧ାଶ ) where ‫ݔ‬ଵ = ܽ, ‫ݔ‬ேାଵ = ܾ
central 1 ݂(ܽ) + ݂(ܾ)
n Error = = ݂(ܽ)(ܾ − ܽ) + [݂(ܾ) − ݂(ܽ)](ܾ − ܽ) = (ܾ − ܽ)
12hଶ 2 2
9.3.2 Composite Trapezoidal ே ேିଵ ௕ ே
ℎ
= ൥݂(ܽ) + 4 ෍ ݂(‫ݔ‬௜ ) + 2 ෍ ݂(‫ݔ‬௜ ) + ݂(ܾ)൩ ℎ = 0.6, ‫ = )݂(ܫ‬න ݂(‫ = ݔ݀)ݔ‬ℎ ෍ ݂(‫ݔ‬௜ ) = 0.6(0.5 + 0.8 + 2) = 1.98
Domain [ܽ, ܾ] is divided into N subintervals by ܰ + 1 points ‫ݔ‬ଵ , ‫ݔ‬ଶ , … , ‫ݔ‬ேାଵ 3 ௔
௜ୀଶ,ସ,଺,… ௜ୀଷ,ହ,଻,… ௜ୀଵ
where ‫ݔ‬ଵ = ܽ, ‫ݔ‬ேାଵ = ܾ
9.4.3 Simpson’s 3/8 method b. trapezoidal
௕
‫ = )݂(ܫ‬න ݂(‫ݔ݀)ݔ‬ ௔ି௕ ேିଵ
௔
Domain [ܽ, ܾ], 4‫ݏݐ݊݅݋݌‬: ‫ݔ‬ଵ = ܽ, ‫ݔ‬ସ = ܾ, ‫ݔ‬ଶ = ܽ + ℎ, ‫ݔ‬ଷ = ܽ + 2ℎ, ℎ = ℎ
ଷ ‫= )݂(ܫ‬ [݂(ܽ) + ݂(ܾ)] + ℎ ෍ ݂(‫ݔ‬௜ )
݂(‫ݔ‬ଵ ) + ݂(‫ݔ‬ଶ ) ݂(‫ݔ‬ଶ ) + ݂(‫ݔ‬ଷ ) ௕ 2
= (‫ݔ‬ଶ − ‫ݔ‬ଵ ) + (‫ݔ‬ଷ − ‫ݔ‬ଶ ) + ⋯ 3ℎ ௜ୀଶ
2 2 ‫ = )݂(ܫ‬න ݂(‫= ݔ݀)ݔ‬ [݂(ܽ) + 3݂(‫ݔ‬ଶ ) + 3݂(‫ݔ‬ଷ ) + ݂(ܾ)] 0.3
݂(‫ݔ‬ே ) + ݂(‫ݔ‬ேାଵ ) ௔ 8 ℎ = 0.3, ‫= )݂(ܫ‬ (0.5 + 4.8) + 0.3(0.6 + ⋯ + 3.2) = 3.165
+ (‫ݔ‬ேାଵ − ‫ݔ‬ே ) 2
2 9.4.4 Composite Simpson’s 3/8 method 0.6
ே ℎ = 0.6, ‫= )݂(ܫ‬ (0.5 + 4.8) + 0.6(0.8 + 2) = 3.27
1 2
= ෍[݂(‫ݔ‬௜ ) + ݂(‫ݔ‬௜ାଵ )](‫ݔ‬௜ାଵ − ‫ݔ‬௜ )] Domain [ܽ, ܾ] is divided into ܰ (ܰ ݅‫ ݕܾ ݈ܾ݁݅ݏ݅ݒ݅݀ ݏ‬3) subintervals with equal c. Simpson’s 3/8
2 ௕ି௔
௜ୀଵ width ℎ = by ܰ + 1 points ‫ݔ‬ଵ , ‫ݔ‬ଶ , … , ‫ݔ‬ேାଵ , where ‫ݔ‬ଵ = ܽ, ‫ݔ‬ேାଵ = ܾ,
ே ேିଵ ேିଶ
When the subintervals have the same width h: 3ℎ
௫ర ௫ళ ௫ಿశభ ‫= )݂(ܫ‬ ൥݂(ܽ) + 3 ෍ [݂(‫ݔ‬௜ ) + ݂(‫ݔ‬௜ାଵ )] + 2 ෍ ݂(‫ݔ‬௜ ) + ݂(ܾ)൩
ே ேିଵ ‫ = )݂(ܫ‬න ݂(‫ ݔ݀)ݔ‬+ න ݂(‫ ݔ݀)ݔ‬+ ⋯ + + න ݂(‫ݔ݀)ݔ‬ 8
௕ ௜ୀଶ,ହ,଼,… ௜ୀସ,଻,ଵ଴,…
ℎ ℎ ௔ ௫ర ௫ಿషమ 3 ∙ 0.3
‫ = )݂(ܫ‬න ݂(‫= ݔ݀)ݔ‬ ෍[݂(‫ݔ‬௜ ) + ݂(‫ݔ‬௜ାଵ )] = [݂(ܽ) + ݂(ܾ)] + ℎ ෍ ݂(‫ݔ‬௜ ) ேିଶ ℎ = 0.3, ‫= )݂(ܫ‬ [0.5 + 3(0.6 + 0.8 + 2 + 3.2) + 2 ∙ 1.3 + 4.8] = 3.116
௔ 2 2 ௫೔శయ
8
௜ୀଵ ௜ୀଶ
= ෍ න ݂(‫ݔ݀)ݔ‬ 3 ∙ 0.6
9.4.1 Simpson’s 1/3 method ௜ୀଵ,ସ,଻,… ௫೔ ℎ = 0.6, ‫= )݂(ܫ‬ [0.5 + 3(0.8 + 2) + 4.8] = 3.0825
8
௔ା௕ ௔ି௕ When the subintervals have the same width h:
Domain [ܽ, ܾ], ‫ݔ‬ଵ = ܽ, ‫ݔ‬ଷ = ܾ, ‫ݔ‬ଶ = ଶ
,ℎ = ଶ
௕
3ℎ
ℎ ‫= )݂(ܫ‬ [݂(‫ݔ‬ଵ ) + 3݂(‫ݔ‬ଶ ) + 3݂(‫ݔ‬ଷ ) + ݂(‫ݔ‬ସ ) + ݂(‫ݔ‬ସ ) + 3݂(‫ݔ‬ହ ) + 3݂(‫) ଺ݔ‬
‫ = )݂(ܫ‬න ݂(‫ݔ(݂[ = ݔ݀)ݔ‬ଵ ) + 4݂(‫ݔ‬ଶ ) + ݂(‫ݔ‬ଷ )] 8
௔ 3 + ݂(‫ ) ଻ݔ‬+ ⋯ + ݂(‫ݔ‬ேିଶ ) + 3݂(‫ݔ‬ேିଵ ) + 3݂(‫ݔ‬ே ) + ݂(‫ݔ‬ேାଵ )]
ℎ ܽ+ܾ ேିଵ ேିଶ
= ൤݂(ܽ) + 4݂ ൬ ൰ + ݂(ܾ)൨ 3ℎ
3 2 = ൥݂(ܽ) + 3 ෍ [݂(‫ݔ‬௜ ) + ݂(‫ݔ‬௜ାଵ )] + 2 ෍ ݂(‫ݔ‬௜ ) + ݂(ܾ)൩
8
௜ୀଶ,ହ,଼,… ௜ୀସ,଻,ଵ଴,…
9.4.2 Composite Simpson’s 1/3 method
ଵ.଼
Domain [ܽ, ܾ] is divided into even number N subintervals with equal width ℎ = Problem 9.1: Compute ∫଴ ݂(‫ ݔ݀)ݔ‬with ℎ = 0.3 ܽ݊݀ ℎ = 0.6
௕ି௔
ே
by ܰ + 1 points ‫ݔ‬ଵ , ‫ݔ‬ଶ , … , ‫ݔ‬ேାଵ , where ‫ݔ‬ଵ = ܽ, ‫ݔ‬ேାଵ = ܾ, x 0 0.3 0.6 0.9 1.2 1.5 1.8
௫య ௫ఱ ௫ಿశభ
f(x) 0.5 0.6 0.8 1.3 2 3.2 4.8
‫ = )݂(ܫ‬න ݂(‫ ݔ݀)ݔ‬+ න ݂(‫ ݔ݀)ݔ‬+ ⋯ + + න ݂(‫ݔ݀)ݔ‬
௔ ௫య ௫ಿషభ
ே ௫೔శభ
a. rectangle
= ෍ න ݂(‫ݔ݀)ݔ‬
ே
௜ୀଶ,ସ,଺,… ௫೔షభ ௕
ℎ = 0.3, ‫ = )݂(ܫ‬න ݂(‫ = ݔ݀)ݔ‬ℎ ෍ ݂(‫ݔ‬௜ ) = 0.3(0.5 + 0.6 + ⋯ + 3.2) = 2.53
ℎ ௔
‫ݔ(݂[ = )݂(ܫ‬ଵ ) + 4݂(‫ݔ‬ଶ ) + ݂(‫ݔ‬ଷ ) + ݂(‫ݔ‬ଷ ) + 4݂(‫ݔ‬ସ ) + ݂(‫ݔ‬ହ ) + ⋯ + ݂(‫ݔ‬ேିଵ ) ௜ୀଵ
3
+ 4݂(‫ݔ‬ே ) + ݂(‫ݔ‬ଷ )]