An equation like dxdy+2y=ex can’t be separated: the y and the ex are added, not multiplied. But it is linear in y, and every linear first-order equation can be solved the same way: multiply through by a carefully chosen function, the integrating factor, and the left side becomes the derivative of a product. This is the last of the exact methods in IB Mathematics AA HL, alongside separation of variables and homogeneous equations.
A first-order differential equation is linear if it can be written in the standard form
dxdy+P(x)y=Q(x)
where P and Q are functions of x only. The coefficient of dxdy must be 1. If it isn’t, divide through first. For example, xdxdy+2y=x2 becomes dxdy+x2y=x, so P(x)=x2 and Q(x)=x.
A tank holds 50 L of pure water. Brine containing 2 g of salt per litre flows in at 3 L/min, and the well-mixed solution flows out at 2 L/min. Let S be the mass of salt (in grams) in the tank after t minutes.
(a) Explain why dtdS=6−50+t2S.
(b) Solve the differential equation, and find the concentration of salt after 50 minutes.
Solution.
(a) Salt flows in at (2 g/L)(3 L/min)=6 g/min. The volume grows by 1 L each minute, so it is 50+t litres, and the concentration is 50+tS g/L. Salt flows out at 50+tS×2 g/min. Rate of change = rate in − rate out.
(b) Standard form: dtdS+50+t2S=6. The integrating factor is
I(t)=e∫50+t2dt=e2ln(50+t)=(50+t)2
So
dtd((50+t)2S)(50+t)2S=6(50+t)2=2(50+t)3+C
At t=0 the water is pure, so S=0: 0=2(50)3+C, giving C=−250000. Then
S=2(50+t)−(50+t)2250000
At t=50: S=2(100)−10000250000=200−25=175 g, in 100 L of solution. The concentration is 1.75 g/L. (In the long run it approaches the inflow concentration of 2 g/L, which makes sense.)
Not dividing by the coefficient of dy/dx first. The integrating factor formula only works when the equation starts dxdy+…. For xdxdy+2y=x2, the P(x) is x2, not 2.
Getting the sign of P wrong. For dxdy−x2y=…, P(x)=−x2 and I(x)=e−2lnx=x−2. Move every y term to the left before reading off P.
Forgetting to multiply the right side by I(x). Both sides get multiplied. The left side becomes dxd(Iy) automatically; the right side becomes I(x)Q(x), which is what you integrate.
Dividing only part of the answer by I(x). From e2xy=31e3x+C, the solution is y=31ex+Ce−2x, not 31ex+C. The constant gets divided too, and that’s what makes the solution curves different shapes.
Not simplifying the integrating factor. Leaving I(x)=e2lnx makes the next step messy. Use log laws: e2lnx=x2.
Adding the constant too late. The +C belongs on the right as soon as you integrate. Adding it after dividing by I(x) gives the wrong family of solutions.
1. (Warm-up) For each equation, state P(x) and Q(x) and find the integrating factor, simplified.
(a) dxdy+3y=x
(b) dxdy−x1y=x2, x>0
(c) dxdy+2xy=x
Solution
(a) P(x)=3, Q(x)=x, I(x)=e3x.
(b) P(x)=−x1, Q(x)=x2, I(x)=e−lnx=x1.
(c) P(x)=2x, Q(x)=x, I(x)=ex2.
2. (Warm-up) Find the general solution of dxdy−y=e2x.
Solution
P(x)=−1, so I(x)=e−x.
dxd(e−xy)=e−xe2x=ex⇒e−xy=ex+C⇒y=e2x+Cex
3. (Warm-up) Find the general solution of dxdy+2xy=2x.
Solution
I(x)=ex2.
dxd(ex2y)=2xex2⇒ex2y=ex2+C⇒y=1+Ce−x2
(The integral ∫2xex2dx=ex2+C uses the substitution u=x2. This equation is also separable; try it that way too.)
4. (Core) Solve xdxdy−2y=x3ex, x>0, given that y=0 when x=1.
Solution
Divide by x: dxdy−x2y=x2ex. Then I(x)=e−2lnx=x−2.
dxd(x2y)=x−2⋅x2ex=ex⇒x2y=ex+C
At (1,0): 0=e+C, so C=−e and
y=x2(ex−e)
5. (Core) Solve dxdy=x−y with y(0)=2, and find y(1) to 3 s.f.
Solution
Rearrange: dxdy+y=x, so I(x)=ex.
dxd(exy)=xex
By parts, ∫xexdx=xex−ex+C. So exy=(x−1)ex+C and y=x−1+Ce−x.
At x=0: 2=−1+C, so C=3 and y=x−1+3e−x. Then y(1)=3e−1≈1.10 (3 s.f.).
(Compare with the rough estimate of 0.75 from two steps of Euler’s method for the same equation.)
6. (Core) Solve cosxdxdy+ysinx=1 for 0≤x<2π, given that y(0)=3.
Solution
Divide by cosx: dxdy+ytanx=secx. As in Example 3, I(x)=secx.
dxd(ysecx)=sec2x⇒ysecx=tanx+C⇒y=sinx+Ccosx
At x=0: 3=0+C. So y=sinx+3cosx.
7. (Core) A pollutant enters a lake, and the mass P kg of pollutant in the lake after t days satisfies
dtdP+0.1P=5e−0.05t,P(0)=0
(a) Find P in terms of t.
(b) Find the maximum mass of pollutant in the lake, and when it occurs (to 3 s.f.).
Solution
(a) I(t)=e0.1t.
dtd(e0.1tP)=5e0.05t⇒e0.1tP=100e0.05t+C
At t=0: 0=100+C, so C=−100 and
P=100(e−0.05t−e−0.1t)
(b) dtdP=100(−0.05e−0.05t+0.1e−0.1t)=0 gives e−0.05t=2e−0.1t, so e0.05t=2 and t=20ln2≈13.9 days. Then e−0.05t=21 and e−0.1t=41, so P=100(21−41)=25 kg. (It is a maximum: P starts at 0, rises, and tends to 0 again.)
8. (Challenge) Solve dxdy+xy=sinx, x>0, given that y=1 when x=π.
Solution
I(x)=elnx=x.
dxd(xy)=xsinx
By parts, with u=x and dv=sinxdx: ∫xsinxdx=−xcosx+∫cosxdx=sinx−xcosx+C. So
xy=sinx−xcosx+C
At (π,1): π=0−π(−1)+C=π+C, so C=0 and
y=xsinx−cosx
9. (Challenge) Solve (x2+1)dxdy+4xy=x, given that y(0)=0, and describe what happens to y as x→∞.
Solution
Divide by x2+1: dxdy+x2+14xy=x2+1x. Then
I(x)=e∫x2+14xdx=e2ln(x2+1)=(x2+1)2
Multiply through:
dxd((x2+1)2y)=x(x2+1)
With u=x2+1, ∫x(x2+1)dx=4(x2+1)2+C. So (x2+1)2y=4(x2+1)2+C, and
y=41+(x2+1)2C
At x=0: 0=41+C, so C=−41 and y=41−4(x2+1)21.
As x→∞, the second term tends to 0, so y→41 (from below).