Integration by Parts, Tables, and Multivariable Calculus Study Guide

Integration by Parts

Integration by parts is a technique derived from the product rule for differentiation. Recall that the product rule is defined as (uv)=uv+uv(uv)' = u'v + uv' or (fg)=fg+fg(f \cdot g)' = f' \cdot g + f \cdot g'. By integrating both sides of this equation, we obtain (uv)dx=uvdx+uvdx\int (uv)' \, dx = \int u'v \, dx + \int uv' \, dx. Solving for the integral of uvuv', we arrive at the standard formula for Integration by Parts: udv=uvvdu\int u \, dv = uv - \int v \, du. For differentiable functions uu and vv, this can also be expressed as f(x)g(x)dx=f(x)g(x)f(x)g(x)dx\int f(x)g'(x) \, dx = f(x)g(x) - \int f'(x)g(x) \, dx.

Consider the example of integrating xexdx\int x e^x \, dx. To solve this, we define u=xu = x and dv=exdxdv = e^x \, dx. This leads to du=1dxdu = 1 \, dx and v=exv = e^x. Applying the integration by parts formula gives xexexdxx \cdot e^x - \int e^x \, dx. Evaluating the remaining integral results in xexex+Cx e^x - e^x + C. This expression can be simplified by factoring to obtain ex(x1)+Ce^x(x - 1) + C.

When choosing which parts of the integrand to assign to uu and dvdv, two general guidelines are provided. First, choose dvdv to be the most complicated part of the integral that can be integrated easily. Second, choose uu so that its derivative uu' is simpler than uu itself. Specific patterns help in choosing these values for common integral forms. For integrals of the form xneaxdx\int x^n e^{ax} \, dx, choose u=xnu = x^n and dv=eaxdxdv = e^{ax} \, dx. For integrals of the form xnln(x)dx\int x^n \ln(x) \, dx, choose u=ln(x)u = \ln(x) and dv=xndxdv = x^n \, dx. For integrals of the form (x+a)(x+b)ndx\int (x+a)(x+b)^n \, dx, choose u=x+au = x+a and dv=(x+b)ndxdv = (x+b)^n \, dx.

In practice problem #12, the integral xln(x)dx\int x \ln(x) \, dx is evaluated by setting u=ln(x)u = \ln(x) and dv=xdxdv = x \, dx. This yields du=1xdxdu = \frac{1}{x} \, dx and v=12x2v = \frac{1}{2}x^2. The solution is 12x2ln(x)12x21xdx\frac{1}{2}x^2 \ln(x) - \int \frac{1}{2}x^2 \cdot \frac{1}{x} \, dx, which simplifies to 12x2ln(x)12xdx\frac{1}{2}x^2 \ln(x) - \frac{1}{2} \int x \, dx, resulting in 12x2ln(x)14x2+C\frac{1}{2}x^2 \ln(x) - \frac{1}{4}x^2 + C. For practice problem #18, (x+2)(x5)5dx\int (x+2)(x-5)^5 \, dx, we set u=x+2u = x+2 and dv=(x5)5dxdv = (x-5)^5 \, dx, leading to du=1dxdu = 1 \, dx and v=16(x5)6v = \frac{1}{6}(x-5)^6. The integration yields 16(x+2)(x5)616(x5)6dx\frac{1}{6}(x+2)(x-5)^6 - \int \frac{1}{6}(x-5)^6 \, dx, resulting in 16(x+2)(x5)6142(x5)7+C\frac{1}{6}(x+2)(x-5)^6 - \frac{1}{42}(x-5)^7 + C. Practice problem #24 integrates (x+1)e3xdx\int (x+1)e^{-3x} \, dx. Setting u=x+1u = x+1 and dv=e3xdxdv = e^{-3x} \, dx results in du=1dxdu = 1 \, dx and v=13e3xv = -\frac{1}{3}e^{-3x}. The solution is 13(x+1)e3x13e3xdx-\frac{1}{3}(x+1)e^{-3x} - \int -\frac{1}{3}e^{-3x} \, dx, which simplifies to 13(x+1)e3x19e3x+C-\frac{1}{3}(x+1)e^{-3x} - \frac{1}{9}e^{-3x} + C.

Practice problem #26 involves (x+1)1/2ln(x+1)dx\int (x+1)^{-1/2} \ln(x+1) \, dx. Let u=ln(x+1)u = \ln(x+1) and dv=(x+1)1/2dxdv = (x+1)^{-1/2} \, dx. Thus du=1x+1dxdu = \frac{1}{x+1} \, dx and v=2(x+1)1/2v = 2(x+1)^{1/2}. The calculation proceeds as 2(x+1)1/2ln(x+1)2(x+1)1/21x+1dx2(x+1)^{1/2} \ln(x+1) - \int 2(x+1)^{1/2} \cdot \frac{1}{x+1} \, dx, which becomes 2x+1ln(x+1)2(x+1)1/2dx2\sqrt{x+1} \ln(x+1) - 2 \int (x+1)^{-1/2} \, dx, resulting in 2x+1ln(x+1)4(x+1)1/2+C2\sqrt{x+1} \ln(x+1) - 4(x+1)^{1/2} + C, or simplified as 2x+1(ln(x+1)2)+C2\sqrt{x+1}(\ln(x+1) - 2) + C. Problem #28 evaluates xx+1dx\int x\sqrt{x+1} \, dx. Setting u=xu = x and dv=(x+1)1/2dxdv = (x+1)^{1/2} \, dx yields du=1dxdu = 1 \, dx and v=23(x+1)3/2v = \frac{2}{3}(x+1)^{3/2}. The solution is 23x(x+1)3/223(x+1)3/2dx\frac{2}{3}x(x+1)^{3/2} - \int \frac{2}{3}(x+1)^{3/2} \, dx, resulting in 23x(x+1)3/2415(x+1)5/2+C\frac{2}{3}x(x+1)^{3/2} - \frac{4}{15}(x+1)^{5/2} + C.

Definite integrals can also be solved using these methods. Practice problem #42 evaluates 12xln(x)dx\int_{1}^{2} x \ln(x) \, dx. Using the previous result for the indefinite integral, 12x2ln(x)14x2\frac{1}{2}x^2 \ln(x) - \frac{1}{4}x^2, we evaluate from 11 to 22. This results in (12(2)2ln(2)14(2)2)(12(1)2ln(1)14(1)2)(\frac{1}{2}(2)^2 \ln(2) - \frac{1}{4}(2)^2) - (\frac{1}{2}(1)^2 \ln(1) - \frac{1}{4}(1)^2), which simplifies to (2ln(2)1)(014)(2 \ln(2) - 1) - (0 - \frac{1}{4}), resulting in a final answer of 2ln(2)342 \ln(2) - \frac{3}{4}.

Integration Using Tables

Standard integral tables, often found in the back of textbooks, provide formulas for complex integrands. The steps for using an integral table involve finding a formula that matches the integrand, identifying the constants such as aa, bb, cc, and dd, and finally using the formula to complete the integration.

In problem #8, the integral is 15x225dx\int \frac{15}{x^2-25} \, dx. We identify the formula 1x2a2dx=12alnxax+a+C\int \frac{1}{x^2 - a^2} \, dx = \frac{1}{2a} \ln |\frac{x-a}{x+a}| + C. Here, a=5a = 5. The calculation is 15×12(5)lnx5x+5+C15 \times \frac{1}{2(5)} \ln |\frac{x-5}{x+5}| + C, which simplifies to 1510lnx5x+5+C\frac{15}{10} \ln |\frac{x-5}{x+5}| + C or 32lnx5x+5+C\frac{3}{2} \ln |\frac{x-5}{x+5}| + C. Problem #21 (listed as #11 in some sources) evaluates 5(x+1)(x+2)dx\int \frac{5}{(x+1)(x+2)} \, dx. Using the formula 1(ax+b)(cx+d)dx=1adbc(lnax+blncx+d)+C\int \frac{1}{(ax+b)(cx+d)} \, dx = \frac{1}{ad-bc} (\ln|ax+b| - \ln|cx+d|) + C, with a=1,b=1,c=1,d=2a=1, b=1, c=1, d=2. The denominator constant becomes 21=12-1 = 1. The result is 5(lnx+2lnx+1)+C5 (\ln|x+2| - \ln|x+1|) + C, which can be written as 5lnx+25lnx+1+C5 \ln |x+2| - 5 \ln |x+1| + C.

Problem #18 involves the integral 64+z2dz\int \frac{6}{\sqrt{4+z^2}} \, dz. Utilizing formula #19 from the table, which fits the form a2+x2xdx\int \frac{\sqrt{a^2+x^2}}{x} \, dx or 1a2+x2dx\int \frac{1}{\sqrt{a^2+x^2}} \, dx. The transcript shows the formula application leading to 4+z22ln2+4+z2z+C\sqrt{4+z^2} - 2 \ln |\frac{2+\sqrt{4+z^2}}{z}| + C. Problem #30 evaluates 19e2tdt\int \frac{1}{9-e^{2t}} \, dt. Let u=etu = e^t and du=etdtdu = e^t \, dt. This transforms the integral to use formula #17: 12(3)ln3+et3et+C\frac{1}{2(3)} \ln |\frac{3+e^t}{3-e^t}| + C, resulting in 16ln3+et3et+C\frac{1}{6} \ln |\frac{3+e^t}{3-e^t}| + C. Problem #34 involves x2x6+1dx\int x^2 \sqrt{x^6+1} \, dx. Let u=x3u = x^3, then du=3x2dxdu = 3x^2 \, dx. The integral becomes 13u2+1du\frac{1}{3} \int \sqrt{u^2+1} \, du, which matches formula #16.

Differential calculus and table usage are applied in real-world scenarios. In problem #67, the marginal cost for a manufacturer is given as MC(x)=1x2+1MC(x) = \frac{1}{\sqrt{x^2+1}} with fixed costs equal to 20002000, meaning C(0)=2000C(0) = 2000. The cost function is the integral of the marginal cost: C(x)=1x2+1dxC(x) = \int \frac{1}{\sqrt{x^2+1}} \, dx. Using formula #18, we find C(x)=ln(x+x2+1)+KC(x) = \ln(x + \sqrt{x^2+1}) + K. Setting C(0)=2000C(0) = 2000, we have 2000=ln(0+02+1)+K2000 = \ln(0 + \sqrt{0^2+1}) + K, which simplifies to 2000=ln(1)+K2000 = \ln(1) + K. Since ln(1)=0\ln(1) = 0, then K=2000K = 2000. The final cost function is C(x)=ln(x+x2+1)+2000C(x) = \ln(x + \sqrt{x^2+1}) + 2000.

Partial Derivatives

A partial derivative is the derivative of a function of two or more variables with respect to one variable, while treating all other variables as constants. The partial derivative of f(x,y)f(x, y) with respect to xx is defined by the limit: fx(x,y)=limh0f(x+h,y)f(x,y)hf_x(x, y) = \lim_{h \to 0} \frac{f(x+h, y) - f(x, y)}{h}. Similarly, the partial derivative with respect to yy is defined as: fy(x,y)=limh0f(x,y+h)f(x,y)hf_y(x, y) = \lim_{h \to 0} \frac{f(x, y+h) - f(x, y)}{h}. Subscript notation is frequently used, where fx(x,y)f_x(x, y) denotes the partial derivative with respect to xx and fy(x,y)f_y(x, y) denotes the partial derivative with respect to yy.

For the function f(x,y)=2x3y5f(x, y) = 2x^3 y^5, the partial derivative with respect to xx is found by treating yy as a constant: fx(x,y)=2(3x2)y5=6x2y5f_x(x, y) = 2(3x^2)y^5 = 6x^2 y^5. The partial derivative with respect to yy is found by treating xx as a constant: fy(x,y)=2x3(5y4)=10x3y4f_y(x, y) = 2x^3 (5y^4) = 10x^3 y^4. In a more complex example, f(x,y)=2x47x3y2xy+2y+5f(x, y) = 2x^4 - 7x^3 y^2 - xy + 2y + 5, the partials are calculated as follows: fx(x,y)=8x37y2(3x2)y+0+0f_x(x, y) = 8x^3 - 7y^2(3x^2) - y + 0 + 0, which simplifies to 8x321x2y2y8x^3 - 21x^2 y^2 - y. For the same function, fy(x,y)=07x3(2y)x+2+0f_y(x, y) = 0 - 7x^3(2y) - x + 2 + 0, which simplifies to 14x3yx+2-14x^3 y - x + 2.

Consider the function involving an exponential: f(x,y)=ex3+y2f(x, y) = e^{x^3 + y^2}. To find the partial derivatives, we use the rule ddu(eu)=euu\frac{d}{du}(e^u) = e^u \cdot u'. Thus, fx(x,y)=ex3+y23x2f_x(x, y) = e^{x^3 + y^2} \cdot 3x^2. Evaluating at the point (1,2)(1, 2), we get fx(1,2)=e13+223(1)2=3e5f_x(1, 2) = e^{1^3 + 2^2} \cdot 3(1)^2 = 3e^5. For the partial with respect to yy, fy(x,y)=ex3+y22yf_y(x, y) = e^{x^3 + y^2} \cdot 2y. Evaluating at the same point gives fy(1,2)=e13+222(2)=4e5f_y(1, 2) = e^{1^3 + 2^2} \cdot 2(2) = 4e^5.

Higher order partial derivatives involve taking subsequent derivatives. For f(x,y)=x3y4e2yf(x, y) = x^3 y^4 - e^{2y}, the first partial with respect to xx is fx=3x2y4f_x = 3x^2 y^4. The second partial fxyf_{xy} (the derivative of fxf_x with respect to yy) is 3x2(4y3)=12x2y33x^2 (4y^3) = 12x^2 y^3. The first partial with respect to yy is fy=4x3y32e2yf_y = 4x^3 y^3 - 2e^{2y}. The derivative fyxf_{yx} (the derivative of fyf_y with respect to xx) is 4(3x2)y3=12x2y34(3x^2)y^3 = 12x^2 y^3. Note that fxy=fyxf_{xy} = f_{yx}. Further derivatives, such as fyxxf_{yxx}, involve differentiating fyxf_{yx} with respect to xx again: fyxx=12(2x)y3=24xy3f_{yxx} = 12(2x)y^3 = 24xy^3.

Optimizing Functions of Several Variables

Optimizing functions of several variables involves finding relative extrema. A function f(x,y)f(x, y) has a relative minimum value at (a,b)(a, b) if the surface curves upward in all directions from that point. It has a relative maximum at (a,b)(a, b) if the surface curves downward in all directions. A saddle point occurs at (a,b)(a, b) if the function has a maximum in one direction and a minimum in another, meaning it is neither a relative maximum nor a relative minimum.

A critical point (a,b)(a, b) is defined as a point where both first-order partial derivatives are zero: fx(a,b)=0f_x(a, b) = 0 and fy(a,b)=0f_y(a, b) = 0. Relative maximum and minimum values can occur only at these critical points.

To find the critical points for the function f(x,y)=2x2+y2+2xy+4x+2y+5f(x, y) = 2x^2 + y^2 + 2xy + 4x + 2y + 5, we first compute the partial derivatives and set them to zero. The derivative with respect to xx is fx=4x+2y+4f_x = 4x + 2y + 4. Setting this to zero gives 4x+2y+4=04x + 2y + 4 = 0. The derivative with respect to yy is fy=2y+2x+2f_y = 2y + 2x + 2. Setting this to zero gives 2(y+x+1)=02(y + x + 1) = 0, which implies x+y+1=0x + y + 1 = 0 or y=1xy = -1 - x. Substituting this expression for yy into the first equation: 4x+2(1x)+4=04x + 2(-1 - x) + 4 = 0. This simplifies to 4x22x+4=04x - 2 - 2x + 4 = 0, leading to 2x+2=02x + 2 = 0, or x=1x = -1. Substituting x=1x = -1 back into the expression for yy gives y=1(1)=0y = -1 - (-1) = 0. Therefore, the critical point is (x,y)=(1,0)(x, y) = (-1, 0).