Double integrals and change of order of integration
Double integrals over plane regions: definitions, properties, Fubini’s theorem, region types, and systematic change of order of integration with worked examples.
Drafted with Aria, reviewed by the AiCanCode.org team · Updated . How we write and check our notes. Spotted an error? Use Give Feedback at the bottom of the page.
Why it matters
Engineers use double integrals to accumulate quantities over a surface or region in the plane: total mass from surface density, total heat over a plate, or total flux through a region. Changing the order of integration often turns a difficult integral into an easy one, which saves time in design calculations and simulations.
Key ideas
Type I (x-first) region: A region R described by a ≤ x ≤ b and g₁(x) ≤ y ≤ g₂(x), where g₁ and g₂ are continuous on [a, b] and g₁(x) ≤ g₂(x) for all x in [a, b]. Diagram to draw: x- and y-axes with vertical strips; draw the curves y = g₁(x) and y = g₂(x); mark x = a, x = b; shade the area between the curves for a ≤ x ≤ b.
Type II (y-first) region: A region R described by c ≤ y ≤ d and h₁(y) ≤ x ≤ h₂(y), where h₁ and h₂ are continuous on [c, d] and h₁(y) ≤ h₂(y). Diagram to draw: axes with horizontal strips; draw x = h₁(y) and x = h₂(y); mark y = c, y = d; shade area between these curves for c ≤ y ≤ d.
Iterated integrals: If R is Type I, the double integral is computed by integrating with respect to y first, then x: ∬R f dA = ∫{x=a}^{b} [∫{y=g₁(x)}^{g₂(x)} f(x, y) dy] dx. If R is Type II, integrate with respect to x first, then y: ∬R f dA = ∫{y=c}^{d} [∫{x=h₁(y)}^{h₂(y)} f(x, y) dx] dy.
Fubini’s Theorem (exact statement for rectangles): Let f be Riemann integrable on the rectangle R = [a, b] × [c, d] (e.g., any continuous f). Then ∬R f(x, y) dA = ∫{a}^{b} [∫{c}^{d} f(x, y) dy] dx = ∫{c}^{d} [∫_{a}^{b} f(x, y) dx] dy. Assumptions commonly used in exams: f is continuous on R (hence Riemann integrable), which ensures the iterated integrals exist and equal the double integral; more generally (in a Lebesgue setting), absolute integrability or nonnegativity also suffice.
Additivity over subregions: If R is a union of a finite number of subregions R₁, R₂, … whose interiors do not overlap (boundaries may meet), then ∬R f dA = ∑ ∬{R_i} f dA. This allows splitting a region to avoid complicated bounds.
Linearity: For scalars α, β and functions f, g integrable on R, ∬_R (αf + βg) dA = α ∬_R f dA + β ∬_R g dA.
Positivity and comparison: If f(x, y) ≥ 0 on R, then ∬_R f dA ≥ 0. If f(x, y) ≤ g(x, y) on R, then ∬_R f dA ≤ ∬_R g dA.
Area via double integral: The area of R is Area(R) = ∬_R 1 dA. Units: m². This is a special case used to check bounds. Geometric check: if f(x, y) ≥ 0, then ∬_R f(x, y) dA equals the volume under z = f(x, y) above R.
Average (mean) value of a function on a region: f_avg = (1/Area(R)) ∬_R f dA, provided Area(R) is finite and nonzero. Units: same as f.
Change of order of integration (concept): Rewriting an iterated integral by interchanging the order of integration, which requires rewriting the limits to describe the same region. Statement used in exams: “If R is a region in the plane for which both orders of iterated integration are valid (i.e., f is integrable over R and R can be represented as both Type I and Type II, possibly by splitting into subregions), then ∬_R f dA = ∫∫ f dx dy = ∫∫ f dy dx.”
How to change the order of integration (algorithm):
- Sketch the region R defined by the given bounds. Draw the bounding curves or lines precisely and shade the region.
- Project R onto the x-axis and the y-axis to determine the full ranges of x and y.
- Decide which order you want (x-first or y-first). Draw thin strips in the direction of the inner integral. If a single pair of bounds cannot describe the whole region, split R into a small number of subregions where simple bounds work.
- Write the new limits carefully from the sketch, ensuring the lower and upper functions are correct for each subregion.
- Verify the new limits describe exactly the same region by checking corner points and intersections.
When to split the region: If a horizontal (or vertical) strip crosses a boundary that changes formula (like switching from one line to another), split at the corresponding y (or x) value so each subregion has simple, single-form bounds.
Note: Change of variables (like polar coordinates) is a separate topic; here we only change the order in Cartesian coordinates. See “Double integrals in polar coordinates” for polar change of variables.
Formulas
Meaning: Sum over subareas R_i with area ΔA_i and sample point (x_i, y_i); limit as partition gets arbitrarily fine. Units: (units of f)·m². Applies when f is Riemann integrable on bounded R.
Type I (x-first): ∬_R f dA = ∫_{x=a}^{b} ∫_{y=g₁(x)}^{g₂(x)} f(x, y) dy dxSymbols: a, b are x-limits (m); g₁, g₂ are lower/upper y-bound curves (m). Applies when R = {(x, y): a ≤ x ≤ b, g₁(x) ≤ y ≤ g₂(x)}.Type II (y-first): ∬_R f dA = ∫_{y=c}^{d} ∫_{x=h₁(y)}^{h₂(y)} f(x, y) dx dySymbols: c, d are y-limits (m); h₁, h₂ are lower/upper x-bound curves (m). Applies when R = {(x, y): c ≤ y ≤ d, h₁(y) ≤ x ≤ h₂(y)}.Fubini’s theorem (rectangle): ∬_R f dA = ∫_{a}^{b} ∫_{c}^{d} f dy dx = ∫_{c}^{d} ∫_{a}^{b} f dx dySymbols: R = [a, b]×[c, d]. Applies if f is continuous on R (or absolutely integrable). Units: (units of f)·m².Additivity: ∬_{R₁∪R₂} f dA = ∬_{R₁} f dA + ∬_{R₂} f dAfor subregions whose interiors are disjoint (boundaries may meet) Applies when R₁ and R₂ have non-overlapping interiors and f is integrable on each.Linearity: ∬_R (αf + βg) dA = α ∬_R f dA + β ∬_R g dAApplies for scalars α, β and integrable f, g.Area of R: Area(R) = ∬_R 1 dAUnits: m². Applies for any measurable region R with finite area.Average value: f_avg = (1/Area(R)) ∬_R f dAUnits: same as f. Applies when Area(R) ≠ 0 and f integrable.Change of order (conceptual equality): ∫_{x=a}^{b} ∫_{y=g₁(x)}^{g₂(x)} f dy dx = ∑ over pieces ∫_{y=c_j}^{d_j} ∫_{x=h_{1j}(y)}^{h_{2j}(y)} f dx dyMeaning: Possibly split R into pieces j so that the new order uses simple bounds in each piece.
Worked examples
Example 1: Direct evaluation over a rectangle (no change of order) Given: Evaluate ∬_R (x + 2y) dA over the rectangle R: 0 ≤ x ≤ 2 m, 1 ≤ y ≤ 3 m.
Steps:
- Recognize R is a rectangle ⇒ either order works. Use y-first then x.
- Write the iterated integral: ∫{x=0}^{2} ∫{y=1}^{3} (x + 2y) dy dx.
- Inner integral (with respect to y): ∫ (x + 2y) dy = x·y + y². Evaluate from y = 1 to y = 3: Value = x(3 − 1) + (3² − 1²) = 2x + 8.
- Outer integral (with respect to x): ∫{0}^{2} (2x + 8) dx = [x² + 8x]{0}^{2} = (4 + 16) − 0 = 20.
Units: x, y in metres ⇒ integrand has units of metres, dA has m² ⇒ result has m³. Final answer: ∬_R (x + 2y) dA = 20 m³.
Diagram to draw (for your rough work): x- and y-axes; draw vertical lines x = 0 and x = 2; horizontal lines y = 1 and y = 3; shade the rectangle.
Example 2: Change order on a triangular region bounded by two lines and an axis Given: Evaluate ∬_R x dA where R is bounded by x = 0, y = x, and y = 2 − x.
- Sketch R: Draw lines y = x (through origin, slope 1) and y = 2 − x (intercepts (0,2) and (2,0)), plus the y-axis x = 0. Intersection of y = x and y = 2 − x is at (1, 1). The triangle has vertices (0, 0), (0, 2), and (1, 1). Shade the area between y = x and y = 2 − x to the right of x = 0.
- Type I description (y as inner variable): For each fixed x from 0 to 1, y runs from y = x up to y = 2 − x. Hence ∬R x dA = ∫{x=0}^{1} ∫_{y=x}^{2−x} x dy dx.
- Evaluate in this order first (to check): Inner integral: ∫{y=x}^{2−x} x dy = x[(2 − x) − x] = 2x − 2x². Outer integral: ∫{0}^{1} (2x − 2x²) dx = [x² − (2/3)x³]_{0}^{1} = 1 − 2/3 = 1/3.
- Change to Type II description (x as inner variable): From the sketch, y ranges from 0 up to 2. For 0 ≤ y ≤ 1, the right boundary is x = y (left boundary x = 0). For 1 ≤ y ≤ 2, the right boundary is x = 2 − y (left boundary x = 0). So ∬R x dA = ∫{y=0}^{1} ∫{x=0}^{y} x dx dy + ∫{y=1}^{2} ∫_{x=0}^{2−y} x dx dy.
- Evaluate: First piece: ∫{0}^{1} [x²/2]{0}^{y} dy = ∫{0}^{1} (y²/2) dy = [y³/6]{0}^{1} = 1/6. Second piece: ∫{1}^{2} [x²/2]{0}^{2−y} dy = ∫{1}^{2} (2−y)²/2 dy. Substitute t = 2 − y ⇒ dt = −dy; as y : 1→2, t : 1→0 ⇒ integral = ∫{0}^{1} t²/2 dt = [t³/6]_{0}^{1} = 1/6. Sum = 1/6 + 1/6 = 1/3.
Units: x in metres ⇒ result has m³. Final answer: ∬_R x dA = 1/3 m³.
Diagram to draw: Axes; line y = x; line y = 2 − x; y-axis x = 0; mark intersection at (1,1); shade the triangle.
Example 3: Change order and evaluate an integral given in one order Given integral: I = ∫{x=0}^{1} ∫{y=0}^{1−x} (x + y) dy dx. Change the order and evaluate.
- Identify R: 0 ≤ x ≤ 1; for each x, 0 ≤ y ≤ 1 − x. This is the right triangle bounded by x = 0, y = 0, and x + y = 1.
- Sketch: Draw axes; line x + y = 1 (intercepts (1,0) and (0,1)); shade the triangle in the first quadrant.
- Change order: For 0 ≤ y ≤ 1, x runs from 0 to 1 − y. So I = ∫{y=0}^{1} ∫{x=0}^{1−y} (x + y) dx dy.
- Evaluate in the original order (quick check): Inner: ∫ (x + y) dy = x y + y²/2 from 0 to 1 − x ⇒ x(1 − x) + (1 − x)²/2 = 1/2 − x²/2. Outer: ∫{0}^{1} (1/2 − x²/2) dx = [x/2 − x³/6]{0}^{1} = 1/2 − 1/6 = 1/3.
- Evaluate in the new order (consistency): Inner: ∫{0}^{1−y} (x + y) dx = [x²/2 + yx]{0}^{1−y} = (1 − y)²/2 + y(1 − y) = 1/2 − y²/2. Outer: ∫{0}^{1} (1/2 − y²/2) dy = [y/2 − y³/6]{0}^{1} = 1/3.
Final answer: I = 1/3 (dimensionless, since no physical units were specified).
Common mistakes
- Forgetting to split the region: Trying to use one pair of bounds when a horizontal (or vertical) strip actually crosses two different boundary curves.
- Dropping the inner variable’s limits during evaluation: After changing order, some still integrate with old limits or with respect to the wrong variable.
- Ignoring units: If x, y are lengths in metres, dA is m²; your answer’s unit is (units of f)·m². Check dimensional consistency.
- Integrating the wrong function: When computing ∫ f dy first, treat x as a constant; when computing ∫ f dx first, treat y as a constant.
- Not checking corner points: New limits should reproduce all vertices/intersections of the original region.
For your exam
Expect questions of these types:
- Definitions/theory: Define double integral (Riemann sum limit), state Fubini’s theorem with assumptions, define Type I and Type II regions, state linearity/additivity and area formula, define average value.
- Short notes: Steps to change the order of integration; when and why to split a region; how to use a sketch to find bounds.
- Derivations: From the Riemann sum on a rectangle to the iterated integral form; proof outline of Fubini’s theorem under continuity (statement-level justification is usually acceptable in first year).
- Numericals: Evaluate double integrals over rectangles/triangles; change order for regions bounded by lines or simple curves; compute area via ∬_R 1 dA; compute average value of f over a region.
- Diagrams: Clean sketch of the region with labeled boundary curves/lines, intercepts, and intersection points. For full marks, indicate the strip direction (horizontal/vertical), write down the corresponding limits, and ensure consistency when multiple subregions are needed.
A full-marks answer typically includes: a clear labeled sketch; correct identification of region type(s); correct and justified limits (including any splits); neat step-by-step integration with the formula written before substituting numbers; units when physical dimensions are given; and a final boxed or clearly highlighted result.
Quick check
- Write the Type II (y-first) limits for the region 0 ≤ x ≤ 2, x ≤ y ≤ 3 − x.
- For R bounded by x = 0, y = 0, and x + y = 4, write the integral for Area(R) as a double integral in both orders.
- True or False: If f is continuous on a rectangle, then ∫{a}^{b} ∫{c}^{d} f dy dx must equal ∫{c}^{d} ∫{a}^{b} f dx dy.
Answers: (1) 0 ≤ y ≤ 3; for 0 ≤ y ≤ 1.5, 0 ≤ x ≤ y; for 1.5 ≤ y ≤ 3, 0 ≤ x ≤ 3 − y. (2) Area = ∫{x=0}^{4} ∫{y=0}^{4−x} 1 dy dx = ∫{y=0}^{4} ∫{x=0}^{4−y} 1 dx dy. (3) True (by Fubini’s theorem under continuity).
Finished this topic? Mark it so your progress, study plan and readiness keep up.
Stuck on something here?