Changing the Order of Integration – Why It Matters
When a double or triple integral is written with a particular order of integration, the limits often reflect the geometry of the region in a way that makes the computation messy or even impossible in closed form. Changing the order of integration rewrites the same integral with the variables integrated in the opposite sequence, usually turning a complicated iterated integral into a much simpler one Worth keeping that in mind. But it adds up..
The technique is a cornerstone of multivariable calculus. It appears in textbooks, research papers, and real‑world applications such as probability (computing joint densities), physics (evaluating fields over volumes), and engineering (stress analysis). Mastering it gives you a powerful tool for tackling integrals that otherwise seem intractable Small thing, real impact..
Below you’ll find a step‑by‑step guide, a concrete example, the underlying theory, and answers to the most common questions students ask.
1. What Does “Change the Order of Integration” Mean?
Consider a double integral over a plane region (R):
[ I=\int_{x=a}^{b}\int_{y=g_1(x)}^{g_2(x)} f(x,y),dy,dx . ]
The order is “(x) first, then (y)”. Changing the order means rewriting the same integral as
[ I=\int_{y=c}^{d}\int_{x=h_1(y)}^{h_2(y)} f(x,y),dx,dy , ]
where the new limits (c,d) and the functions (h_1(y),h_2(y)) describe the same region (R) but with the roles of (x) and (y) swapped.
For triple integrals the idea is identical: you permute the three variables, adjusting the limits accordingly.
2. When Should You Switch the Order?
| Situation | Reason to switch |
|---|---|
| The inner integral is impossible to evaluate in closed form. | |
| You need to apply Fubini’s theorem to justify interchange. | |
| The region is easier to describe with the opposite variable as the outer one. | |
| The integrand contains a factor that depends only on the inner variable. | Guarantees equality of the two iterated integrals. |
If you encounter any of the above, try reordering before attempting heavy algebra.
3. Step‑by‑Step Procedure
-
Sketch the region (R).
- Plot the curves that define the boundaries.
- Identify the “left‑most” and “right‑most” (or “bottom‑most” and “top‑most”) points.
-
Write the original limits in the form
[ a\le x\le b,\qquad g_1(x)\le y\le g_2(x). ] -
Find the range of the new outer variable Less friction, more output..
- For a switch to (dy,dx) → (dx,dy), determine the smallest and largest values of (y) that occur in (R). Call them (c) and (d).
-
For each fixed (y) in ([c,d]), determine the corresponding (x)-interval.
- Solve the boundary equations for (x) in terms of (y).
- The left boundary becomes (x = h_1(y)) and the right boundary becomes (x = h_2(y)).
-
Rewrite the integral with the new order:
[ I = \int_{y=c}^{d}\int_{x=h_1(y)}^{h_2(y)} f(x,y),dx,dy . ] -
Evaluate the new iterated integral, starting with the inner integral (now in (x)) and then the outer one.
4. Worked Example
Evaluate
[ I=\int_{0}^{1}\int_{y}^{\sqrt{y}} e^{x^2},dx,dy . ]
4.1 Sketch the region
The limits tell us:
- (y) runs from (0) to (1).
- For a fixed (y), (x) runs from the line (x=y) up to the curve (x=\sqrt{y}) (or (y=x^2)).
The region is bounded by:
- the line (x=y) (a 45° line),
- the parabola (y=x^2) (opening upward),
- and the vertical lines (x=0) and (x=1) (the intersection points of the curves).
Plotting these curves shows a “curved triangle” with vertices at ((0,0)), ((1,1)), and ((1,0)) Less friction, more output..
4.2 Determine the new limits
We want to integrate (x) first, then (y).
- The smallest (x) in the region is (0); the largest is (1). So (x\in[0,1]).
- For a fixed (x), the lower bound on (y) comes from the parabola (y=x^2); the upper bound comes from the line (y=x).
Thus
[ 0\le x\le 1,\qquad x^2\le y\le x . ]
4.3 Rewrite the integral
[ I=\int_{x=0}^{1}\int_{y=x^2}^{x} e^{x^2},dy,dx . ]
4.4 Evaluate
The inner integral is trivial because the integrand does not depend on (y):
[ \int_{y=x^2}^{x} e^{x^2},dy = e^{x^2},(x - x^2). ]
Now integrate with respect to (x):
[ I = \int_{0}^{1} e^{x^2}(x - x^2),dx = \int_{0}^{1} x e^{x^2},dx - \int_{0}^{1} x^2 e^{x^2},dx . ]
The first term is solved with the substitution (u=x^2), (du=2x,dx):
[ \int_{0}^{1} x e^{x^2},dx = \frac12\int_{0}^{1} e^{u},du = \frac12(e-1). ]
For the second term, integrate by parts or use the known antiderivative (\int x^2 e^{x^2}dx = \frac12 x e^{x^2} - \frac14\sqrt{\pi},\text{erfi}(x)). Evaluating from 0 to 1 gives
[ \int_{0}^{1} x^2 e^{x^2},dx = \frac12 e - \frac{\sqrt{\pi}}{4},\text{erfi}(1). ]
Putting the pieces together:
[ I = \frac12(e-1) - \left(\frac12 e - \frac{\sqrt{\pi}}{4},\text{erfi}(1)\right) = \frac{\sqrt{\pi}}{4},\text{erfi}(1) - \frac12 . ]
Thus, by swapping the order we turned an impossible‑looking integral into a combination of elementary and special‑function terms.
5. Theoretical Justification – Fubini’s Theorem
5. Theoretical Justification – Fubini’s Theorem
The ability to interchange the order of integration in double integrals is grounded in Fubini’s Theorem, a cornerstone of multivariable calculus. This theorem states that if
5. Theoretical Justification – Fubini’s Theorem
The ability to interchange the order of integration in double integrals is grounded in Fubini’s Theorem, a cornerstone of multivariable calculus. This theorem states that if a function (f(x,y)) is integrable over a rectangle (R=[a,b]\times[c,d]) (or over a more general measurable set (D\subset\mathbb{R}^{2})) and if the double integral of its absolute value is finite,
[ \iint_{D} |f(x,y)|,dA < \infty , ]
then the integral can be computed as an iterated integral in either order:
[ \iint_{D} f(x,y),dA = \int_{a}^{b}!Worth adding: ! !\int_{c}^{d} f(x,y),dy,dx = \int_{c}^{d}!\int_{a}^{b} f(x,y),dx,dy .
In practice, the theorem guarantees that the two “slices” of the region—horizontal slices (integrating in (y) first) and vertical slices (integrating in (x) first)—are equivalent in measure and that the order of integration does not affect the final value. The proof hinges on approximating (f) by simple functions and applying Tonelli’s theorem for non‑negative functions, followed by a limiting argument for integrable functions.
Real talk — this step gets skipped all the time.
5.1 When the Theorem Fails
If the function is not absolutely integrable (for instance, if it has an improper singularity that is only conditionally convergent), then swapping the order can change the value or even lead to divergence. In such cases, one must be careful and often resort to principal value integrals or regularization techniques. A classic example is the double integral of (\frac{\sin(xy)}{xy}) over the entire plane, which converges in one order but diverges in the other.
5.2 Practical Checklist
| Step | What to Verify | Why it Matters |
|---|---|---|
| 1. | Absolute integrability | Allows safe interchange of limits and integrals. Here's the thing — |
| 2. | ||
| 5. On top of that, | ||
| 3. | ||
| 4. So | Finite area of (D) | Prevents “infinite strip” problems. In practice, |
Following this checklist virtually eliminates the risk of errors when re‑ordering an iterated integral And that's really what it comes down to..
6. Common Pitfalls and How to Avoid Them
| Pitfall | Example | Fix |
|---|---|---|
| Confusing the roles of (x) and (y) | Writing (x^2\le y\le x) when the correct bounds are (y^2\le x\le y). | Redraw the region, list the curves explicitly, then check the inequalities by plugging in sample points. |
| Ignoring the intersection points | Assuming (x) runs from 0 to 1 when the curves intersect at (x=\frac12). Which means | Solve the equations (y=x) and (y=x^2) simultaneously to find exact limits. |
| Overlooking absolute integrability | Swapping order for (\int_{-1}^{1}\int_{0}^{\infty}\frac{x}{x^2+y^2},dy,dx). | Verify that (\iint |
| Misapplying Fubini on unbounded sets | Integrating over ([0,\infty)\times[0,1]) without checking convergence. | Test convergence of one iterated integral before swapping. |
7. Advanced Extensions
7.1 Triple Integrals
For a triple integral over a region (E\subset\mathbb{R}^{3}), the same principle applies:
[ \iiint_{E} f(x,y,z),dV = \int_{a}^{b}!!\int_{c}^{d}!!\int_{e}^{f} f(x,y,z),dz,dy,dx , ]
provided (f) is absolutely integrable over (E). The process of changing the order involves describing the bounds for two variables in terms of the third, then iterating accordingly.
7.2 Cylindrical and Spherical Coordinates
When the region has radial symmetry, it is often advantageous to switch to cylindrical ((r,\theta,z)) or spherical ((\rho,\phi,\theta)) coordinates. The Jacobian determinant (e.g., (r) for cylindrical, (\rho^2\sin\phi) for spherical) must be included in the integrand. The same care with bounds and integrability applies.
8. Conclusion
Re‑ordering an iterated integral is a powerful technique that turns an otherwise intractable problem into a manageable one. By carefully sketching the region, translating geometric boundaries into algebraic inequalities, and verifying the conditions of Fubini’s Theorem, you can confidently swap the order of integration and simplify the evaluation. Whether you are working with elementary functions or more exotic integrands involving special functions, the same systematic approach—draw, bound, rewrite, integrate—remains valid.
Remember that the key to success lies in visual intuition and rigorous verification. A well‑drawn diagram often reveals the hidden simplicity of the region, while a thorough check of integrability ensures mathematical correctness. Armed with these tools, you can tackle a wide array of double, triple, and higher‑dimensional integrals with confidence and elegance.
