Example 29 Given x₁ = x₀ + h, x₂ = x₀ + 2h, x₃ = x₀ + 3h (h > 0) and u₃(x) = (x - x₀)(x - x₁)(x - x₂)(x - x₃),

Prove that Δ³u₃(x) = 4! h³ (x - x₀).

Solution: Δu₃ = u₃(x + u) - u₃(x) = (x + h - x₀)(x + h - x₁)(x + h - x₂)(x + h - x₃)- (x - x₀)(x - x₁)(x - x₂)(x - x₃)=
(x + h - x₀)(x - x₀)(x - x₁)(x - x₂) - (x - x₀)(x - x₁)(x - x₂)= (x - x₀)(x - x₁)(x - x₂) [(x + h - x₀) - (x - x₃)]= 4h(x -
x₀)(x - x₁)(x - x₂) [∵ x₃ = x₀ + 3h]; Δ²u₃(x) = ΔΔu₃(x) = 4h·3h [(x + h - x₀)(x + h - x₁) - (x - x₀)(x - x₁)] =
4h·3h [(x - x₀)(x + h - x₀ - x + x₁)] = 4h·3h·2h (x - x₀) = 4! h³ (x - x₀) [∵ x₁ = x₀ + h]
Example 10 Verify that, for f(x) = 5x + 6, the values of the integral ∫[a to a+2h] f(x) dx obtained by
Simpson's one-third method and by Trapezoidal rule with h as step length are equal. Give reasons
for this equality.
Solution: Here, f(x) = 5x + 6, x₀ = a, x₁ = a + h, x₂ = a + 2h. Then, y₀ = 5a + 6, y₁ = 5a + 5h + 6, y₂ = 5a +
10h + 6. Now, by Simpson’s one-third rule, the value of the integral is given by Iₛ = (h/3) [y₀ + 4y₁ +
y₂] = (h/3) [5a + 6 + 20a + 20h + 24 + 5a + 10h + 6] = 2h (5a + 5h + 6). Also by Trapezoidal rule, the
value of the integral is Iₜ = (h/2) [y₀ + 2y₁ + y₂] = (h/2) [5a + 6 + 10a + 10h + 12 + 5a + 10h + 6] = 2h
(5a + 5h + 6). Thus, it is seen that Iₛ = Iₜ in this case. The reason for this equality is the same as
discussed in Ex. 4.
Example 20. Write down the quadratic polynomial which takes the same value as f(x) at x = -1, 0,
1 and integrate it to obtain the integration rule ∫ from -1 to 1 f(x) dx = 1/3 [f(-1) + 4f(0) + f(1)]
Assuming the error to have the form Af⁽⁴⁾(ξ), -1 < ξ < 1, find the value of A.
Solution: Since f(x) is known only for three values of x, i.e., for x = -1, 0, and 1, so the second order
differences are constant and third and higher order differences are not possible. Here, x₀ = a = -1
and the length of each interval (h) = 1. Now, f(x) = f(x₀ + uh)= f{-1 + (x + 1)·1} [Since x = x₀ + hu = -1 +
1·u, so, u = x + 1]= E^(1+x) f(-1)= (1 + Δ)^(1+x) f(-1) = [1 + (1 + x)Δ + ((1 + x)x / 2!) Δ²] f(-1)= f(-1) + (1 +
x)Δf(-1) + (x² + x)/2 Δ²f(-1) = f(-1) + (1 + x){f(0) - f(-1)} + (x² + x)/2 {f(1) - 2f(0) + f(-1)}= x²[1/2 f(1) - f(0) +
1/2 f(-1)] + x[1/2 f(1) - 1/2 f(-1)] + f(0) which is the required quadratic polynomial. Now, ∫ from -1 to
1 f(x) dx = [1/2 f(-1) - f(0) + 1/2 f(1)] [x³/3] from -1 to 1 + [1/2 f(1) - 1/2 f(-1)] [x²/2] from -1 to 1 + f(0)[x]
from -1 to 1= 1/3 [f(-1) + 4f(0) + f(1)] after simplification Again, Error = ∫ from x₀ to xₙ Rₙ₊₁(x) dx = ∫
from x₀ to xₙ [ω(x) f⁽ⁿ⁺¹⁾(ξ) / (n + 1)!] dx = ∫ from -1 to 1 [(x + 1)(x - 0)(x - 1)(x - 2)/4!] f⁽⁴⁾(ξ) dx [Since
the error is given in the form Af⁽⁴⁾(ξ)] = f⁽⁴⁾(ξ)/4! ∫ from -1 to 1 (x⁴ - 2x³ - x² + 2x) dx= -1/90 f⁽⁴⁾(ξ)Now,
comparing with the given result, A = -1/90
Example 25. By integrating the Lagrangian three point formula when the abscissas are at equal
spacing h, with the origin taken at the central point, obtain the formula. ∫_{-h}^{x} f(x) dx = (5h³ -
3hx² + 2x³)/(12h²) * f(-h) + (2h³ + 3h²x - x³)/(3h²) * f(0) - (h³ - 3hx² - 2x³)/(12h²) * f(h) + E(x) and show
that the truncation error is expressible in the form: E(x) = ∫_{-h}^{x} (x² - h²) * f[-h, 0, h, x] dx
Solution: The three-point Lagrangian formula for interpolation for the points x₀, x₁, x₂ is f(x) = [(x -
x₁)(x - x₂)/((x₀ - x₁)(x₀ - x₂))] * f(x₀) + [(x - x₀)(x - x₂)/((x₁ - x₀)(x₁ - x₂))] * f(x₁) + [(x - x₀)(x - x₁)/((x₂ - x₀)(x₂ -
x₁))] * f(x₂) + E₁(x) where E₁(x) = (x - x₀)(x - x₁)(x - x₂) * f[x₀, x₁, x₂, x] If we take x₀ = -h, x₁ = 0, x₂ = h, then
f(x) = x(x - h)/(2h²) * f(-h) - (x² - h²)/h² * f(0) + x(x + h)/(2h²) * f(h) + E₁(x) where E₁(x) = (x + h)x(x - h) *
f[-h, 0, h, x] = x(x² - h²) * f[-h, 0, h, x] Now, integrating both sides with respect to x from -h to x, we
get ∫_{-h}^{x} f(x) dx = f(-h)/(2h²) * ∫_{-h}^{x} (x² - xh) dx - f(0)/h² * ∫_{-h}^{x} (x² - h²) dx + f(h)/(2h²) *
∫_{-h}^{x} (x² + xh) dx + E(x) where E(x) = ∫_{-h}^{x} E₁(x) dx = ∫_{-h}^{x} (x + h)x(x - h) * f[-h, 0, h, x] dx
Also, ∫_{-h}^{x} (x² - xh) dx = [x³/3 - hx²/2]_{-h}^{x} = (2x³ - 3hx² + 5h³)/6 ∫_{-h}^{x} (x² - h²) dx = (2h³ +
3h²x - x³)/3 ∫_{-h}^{x} (x² + xh) dx = (2x³ + 3hx² - h³)/6 Thus, ∫_{-h}^{x} f(x) dx = (5h³ - 3hx² + 2x³)/(12h²)
* f(-h) + (2h³ + 3h²x - x³)/(3h²) * f(0) - (h³ - 3hx² - 2x³)/(12h²) * f(h) + E(x)where E(x) = ∫_{-h}^{x} x(x² -
h²) * f[-h, 0, h, x] dx
Example 3.4.17 A certain function f, defined on the interval (0, 1) is such that f(0) = 0, f(1/2) = -1,
f(1) = 0. Find the quadratic polynomial p(x) which agrees with f(x) for x = 0, 1/2, 1. If |d³f/dx³| ≤ 1 for
0 ≤ x ≤ 1, show that |f(x) − p(x)| ≤ 1/12 for 0 ≤ x ≤ 1.
Solution: Let the quadratic polynomial p(x) be p(x) = p₀ + p₁x + p₂x². Thus, we have p(0) = f(0) = 0
⇒ p₀ = 0 p(1/2) = f(1/2) = -1 ⇒ p₀ + p₁/2 + p₂/4 = -1; p(1) = f(1) = 0 ⇒ p₀ + p₁ + p₂ = 0
Solving for p₀, p₁, p₂ we get, p₀ = 0, p₁ = -4, p₂ = 4. ∴ p(x) = -4x + 4x². Thus, the remainder in
approximating f(x) by p(x) is R₃(x) = |f(x) − p(x)| = |(x − 0)(x − 1/2)(x − 1)| * |f‴(ξ)| / 3! ≤ |(x)(x − 1/2)(x −
1)| / 6 (as |f‴(ξ)| ≤ 1) ∴ R₃(x) = |f(x) − p(x)| ≤ |x(x − 1)(2x − 1)| / 12. Now for 0 ≤ x ≤ 1,|x(x − 1)| = |x(1 −
x)| = 1/4 − 1/4 + x − x² = 1/4 * (x − 1/2)² ≤ 1/4 < 1 Again as 0 ≤ x ≤ 1 ⇒ 0 ≤ 2x ≤ 2 ⇒ −1 ≤ 2x − 1 ≤ 1 ∴ |2x
− 1| ≤ 1. Thus, |f(x) − p(x)| < 1/12.
Example 4.4.10 Use the Stirling’s interpolation formula to obtain the result: d/dx [f(x)] = (2/3)[f(x +
1) – f(x – 1)] – (1/12)[f(x + 2) – f(x – 2)] considering the differences up to third order.
Solution: The Stirling’s formula is: f(x) = y₀ + x(Δy₀ + Δy₋₁)/2 + x²/2! * Δ²y₋₁ + x(x² – 1)/3! * (Δ³y₋₁ +
Δ³y₋₂)/2 + ...Differentiating w.r.t. x, we get:d/dx[f(x)] = (Δy₀ + Δy₋₁)/2 + x * Δ²y₋₁ + (3x² – 1)/6 * (Δ³y₋₁ +
Δ³y₋₂)/2 Now putting x = 0, we get: d/dx[y(0)] = (Δy₀ + Δy₋₁)/2 – 1/12(Δ³y₋₁ + Δ³y₋₂) Substituting, Δy₀
= y₁ – y₀ ; Δy₋₁ = y₀ – y₋₁ :: Δ³y₋₁ = y₂ – 3y₁ + 3y₀ – y₋₁::Δ³y₋₂ = y₁ – 3y₀ + 3y₋₁ – y₋₂ and then simplifying,
we get: d/dx[f(x₀)] = (2/3)(y₁ – y₋₁) – (1/12)(y₂ – y₋₂) Now shifting the origin at x, we get: d/dx[f(x)] =
(2/3)[f(x + 1) – f(x – 1)] – (1/12)[f(x + 2) – f(x – 2)]
Example 5.14.26 If f(x) is a quadratic polynomial, deduce that ∫₁³ f(x) dx = 1/12 [f(0) + 22f(2) + f(4)]
Solution: As f(x) is a quadratic polynomial, the third and higher order differences are zero. Again
we have to establish an integration formula in terms of f(0), f(2) and f(4).So we take x₀ = a = 0, and
interval of differencing h = 2. Now ∫₁³ f(x) dx = ∫₁³ f[0 + x/2 · 2] dx = ∫₁³ [1 + Δ^{x/2}f(0)] dx = ∫₁³ [1 +
x/2 · Δ + x/2(x/2 – 1)/2! · Δ²] f(0) dx = ∫₁³ f(0) + x · Δf(0)/2 + (x² – 2x)/8 · Δ²f(0) dx = f(0)|₁³ + (Δf(0)/2) ∫₁³ x
dx + (Δ²f(0)/8) ∫₁³ (x² – 2x) dx = f(0) + (Δf(0)/4) · 8 + (Δ²f(0)/8) · [2/3 · x³]₁³ = 1/12 [24f(0) + 24Δf(0) +
Δ²f(0)] = 1/12 [24f(0) + 24{f(2) – f(0)} + {f(4) – 2f(2) + f(0)}] = 1/12 [24f(2) + f(4) – 2f(2) + f(0)] = 1/12
[f(0) + 22f(2) + f(4)]
Q)Obtain the approximate quadrature formula ∫₀³ⁿ f(x) dx ≈ n [3/8 f(0) + 1/24 {19f(n) - 5f(2n) +
f(3n)}]Solution: Here f(x) is known for four values of x, x = 0, n, 2n, 3n. Thus third order differences
are constant. Here x₀ = a = 0, h = n; ∫₀ⁿ f(x) dx = ∫₀ⁿ f[0 + (x/n) * n] dx = ∫₀ⁿ (1 + Δ)^(x/n) f(0) dx = ∫₀ⁿ
[1 + (x/n)Δ + (x/n)((x/n) - 1)/2 Δ² + (x/n)((x/n) - 1)((x/n) - 2)/6 Δ³] f(0) dx = ∫₀ⁿ [f(0) + xΔf(0)/n + (x² -
nx)/2n² Δ²f(0) + (x³ - 3nx² + 2n²x)/6n³ Δ³f(0)] dx = f(0)x|₀ⁿ + Δf(0)/n x²/2|₀ⁿ + Δ²f(0)/2n² (x³ - nx²)/3|₀ⁿ +
Δ³f(0)/6n³ [x⁴/4 - 3nx³/3 + 2n²x²/2]|₀ⁿ = n [f(0) + Δf(0)/2 - Δ²f(0)/12 + Δ³f(0)/24]
= n [f(0) + 1/2(f(n) - f(0)) - 1/12(f(2n) - 2f(n) + f(0)) + 1/24(f(3n) - 3f(2n) + 3f(n) - f(0))] = n [3/8 f(0) +
1/24 {19f(n) - 5f(2n) + f(3n)}]
Q)Prove the approximate quadrature formula, for a polynomial f(x) of degree 4. ∫₀⁴ f(x) dx = (2/45)
[7{f(0) + f(4)} + 32{f(1) + f(3)} + 12f(2)]Solution: Here the degree of f(x) is 4. Therefore, fourth order
differences are constants and the function f(x) is known for five points 0, 1, 2, 3, 4. If we shift the
origin to x = -2, we have to prove ∫₋²² f(x) dx = (2/45) [7{f(-2) + f(2)} + 32{f(-1) + f(1)} + 12f(0)] to prove
the required formula. Here x₀ = a = -2, h = 1, ∴ ∫₋²² f(x) dx = ∫₋²² [-2 + (x + 2)·1] dx = ∫₋²² (1 + Δ)^(x+2)
f(-2) dx = ∫₋²² [1 + (x + 2)Δ + ((x + 2)(x + 1)/2) Δ² + ((x + 2)(x + 1)x/6) Δ³ + ((x + 2)(x + 1)x(x - 1)/24) Δ⁴] f(-
2) dx = f(-2)x|₋²² + Δf(-2) (x²/2 + 2x)|₋²² + Δ²f(-2)/2 (x³/3 + 3x²/2 + 2x)|₋²² + Δ³f(-2)/6 (x⁴/4 + x³ + x²)|₋²² +
Δ⁴f(-2)/24 (x⁵/5 + x⁴/2 + x³/3 + x²)|₋²² = 4f(-2) + 8Δf(-2) + 20/3 Δ²f(-2) + 8/3 Δ³f(-2) + 14/45 Δ⁴f(-2) = 4f(-
2) + 8[f(-1) - f(-2)] + 20/3 [f(0) - 2f(-1) + f(-2)] + 8/3 [f(1) - 3f(0) + 3f(-1) - f(-2)] + 14/45 [f(2) - 4f(1) +
6f(0) - 4f(-1) + f(-2)] = (4 - 8 + 20/3 - 8/3 + 14/45) f(-2) + (8 - 40/3 + 8 + 56/45) f(-1) + (20/3 - 8 + 84/5)
f(0) + (8/3 - 56/45) f(1) + 14/45 f(2) = 14/45 f(-2) + 64/45 f(-1) + 24/45 f(0) + 64/45 f(1) + 14/45 f(2) =
(2/45) [7{f(-2) + f(2)} + 32{f(-1) + f(1)} + 12f(0)]
Example 5.14.30: If f(x) is a polynomial of degree 2, prove that ∫[0 to 1] f(x) dx = (1/12)[5f(0) + 8f(1)
- f(2)] Solution: If we shift the origin to -1, we have to prove: ∫[-1 to 0] f(x) dx = (1/12)[5f(-1) + 8f(0) -
f(1)] Since f(x) is a polynomial of degree 2, then the second order differences are constants. Here
x₀ = a = -1, h = 1, ∴ ∫[-1 to 0] f(x) dx = ∫[-1 to 0] f[-1 + (x + 1)·1] dx = ∫[-1 to 0] (1 + Δ)^{x+1} f(-1) dx = ∫[-
1 to 0] [1 + (x + 1)·Δ + ((x + 1)·x)/2 · Δ²] f(-1) dx = f(-1)·|x| from -1 to 0 + Δf(-1)·|x²/2 + x| from -1 to 0 +
(Δ²f(-1)/2)·|x³/3 + x²/2| from -1 to 0= f(-1) + Δf(-1)[-1/2 + 1] + (Δ²f(-1)/2)[1/3 - 1/2] = f(-1) + (1/2)[f(0) -
f(-1)] - (1/12)[f(1) - 2f(0) + f(-1)] = (1 - 1/2 - 1/12)f(-1) + (1/2 + 1/6)f(0) - (1/12)f(1) = (5/12)f(-1) +
(8/12)f(0) - (1/12)f(1= (1/12)[5f(-1) + 8f(0) - f(1)]