The document describes the trapezoidal rule and composite trapezoidal rule for numerical integration. The trapezoidal rule approximates the area under a curve as a trapezoid and has an error term related to the second derivative. The composite trapezoidal rule divides the region into panels and applies the trapezoidal rule to each, adding the areas to approximate the total integral. It also introduces the recursive trapezoidal rule which efficiently updates the integral calculation when the number of panels is doubled.
The document describes the trapezoidal rule and composite trapezoidal rule for numerical integration. The trapezoidal rule approximates the area under a curve as a trapezoid and has an error term related to the second derivative. The composite trapezoidal rule divides the region into panels and applies the trapezoidal rule to each, adding the areas to approximate the total integral. It also introduces the recursive trapezoidal rule which efficiently updates the integral calculation when the number of panels is doubled.
The document describes the trapezoidal rule and composite trapezoidal rule for numerical integration. The trapezoidal rule approximates the area under a curve as a trapezoid and has an error term related to the second derivative. The composite trapezoidal rule divides the region into panels and applies the trapezoidal rule to each, adding the areas to approximate the total integral. It also introduces the recursive trapezoidal rule which efficiently updates the integral calculation when the number of panels is doubled.
The document describes the trapezoidal rule and composite trapezoidal rule for numerical integration. The trapezoidal rule approximates the area under a curve as a trapezoid and has an error term related to the second derivative. The composite trapezoidal rule divides the region into panels and applies the trapezoidal rule to each, adding the areas to approximate the total integral. It also introduces the recursive trapezoidal rule which efficiently updates the integral calculation when the number of panels is doubled.
If n = 1 (one panel), as illustrated in Figure 6.2, we have l0 = (x − x1)/(x0 − x1) = −(x − b)/ h. Therefore,
Also l1 = (x − x0)/(x1 − x0) = (x − a)/ h, so that
Substitution in Eq. (6.2a) yields
(6.3)
which is known as the trapezoidal rule. It represents the area of the trapezoid in Figure 6.2.
The error in the trapezoidal rule
is the area of the region between f(x) and the straight-line interpolant, as indicated in Figure 6.2. It can be obtained by integrating the interpolation error in Eq. (3.3):
(6.4)
Composite Trapezoidal Rule
Figure 6.3. Composite trapezoidal rule.
In practice, the trapezoidal rule is applied in a piecewise fashion. Figure 6.3 shows the region (a, b) divided into n panels, each of width h. The function f(x) to be integrated is approximated by a straight line in each panel. From the trapezoidal rule, we obtain for the approximate area of a typical (ith) panel,
b Hence the total area, representing ∫ f (x) dx, is
(6.5)
which is the composite trapezoidal rule.
The truncation error in the area of a panel is – see Eq. (6.4) —
where ξi lies in (xi, xi+1). Hence the truncation error in Eq. (6.5) is
(a)
But
where f¯'' is the arithmetic mean of the second derivatives. If f''(x) is continuous, there must be a point ξ in (a, b) at which f''(ξ) = f¯'', enabling us to write
Therefore, Eq. (a) becomes
(6.6)
It would be incorrect to conclude from Eq. (6.6) that E = ch2 (c being a constant), because f''(ξ) is not entirely independent of h. A deeper analysis of the error shows that if f(x) and its derivatives are finite in (a, b), then
E = c1h2 + c2h4 + c3h6 + . . .
(6.7)
Recursive Trapezoidal Rule
Let Ik be the integral evaluated with the composite trapezoidal rule using 2k−1 panels. Note that if k is increased by one, the number of panels is doubled. Using the notation
H=b−a
Eq. (6.5) yields the following results for k = 1, 2, and 3.
k = 1 (one panel):
I1 = [f(a) + f(b)] H/2
(6.8)
k = 2 (two panels):
We can now see that for arbitrary k>1 we have
(6.9a)
which is the recursive trapezoidal rule. Observe that the summation contains only the new nodes that were created when the number of panels was doubled. Therefore, the computation of the sequence I1, I2, I3, . . . , Ik from Eqs. (6.8) and (6.9) involves the same amount of algebra as the calculation of Ik directly from Eq. (6.5). The advantage of using the recursive trapezoidal rule is that it allows us to monitor convergence and terminate the process when the difference between Ik−1 and Ik becomes sufficiently small. A form of Eq. (6.9a) that is easier to remember is
(6.9b) where h = H/n is the width of each panel.
Illustrative Example:
π Evaluate the bounds on ∫ sin(x)dx with the composite trapezoidal rule using (1) 8 panels; and (2) 16 panels.
Solution of Part (1). With eight panels there are nine nodes spaced at h = π/8. The abscissas of the nodes are xi = iπ/8, i = 0, 1, . . . , 8. From Eq. (6.5) we get
The error is given by Eq. (6.6):
where 0 < ξ < π. Because we do not know the value of ξ, we cannot evaluate E, but we can determine its bounds:
π Therefore, I + Emin < ∫ sin(x)dx < I + Emax, or
0 The exact integral is, of course, 2.
Solution of Part (2). The new nodes created by the doubling of panels are located at the midpoints of the old panels. Their abscissas are
