540 CHAPTER 29. FIRST ORDER SCALAR ODE

2. Multiply by exp(A(t))

chain rule and product rule︷ ︸︸ ︷exp(A(t))

(y′+a(t)y

)=

ddt

(exp(A(t))y) = exp(A(t))b(t)

3. Do∫

to both sides. Pick F (t) ∈∫

exp(A(t))b(t)dt.

exp(A(t))y(t) = F (t)+C

y(t) = exp(−A(t))F (t)+C exp(−A(t))

This proves the following theorem.

Theorem 29.1.7 The solutions to the equation, y′+ a(t)y = b(t) consist of all functionsof the form y(t) = e−A(t)F (t)+ e−A(t)C where F (t) ∈

∫eA(t)b(t)dt and C is a constant,

A′ (t) = a(t).

Finally, here is a uniqueness theorem.

Theorem 29.1.8 If a(t) is a continuous function, there is at most one solution to the initialvalue problem, y′+a(t)y = b(t) , y(r) = y0.

Proof: If there were two solutions y1 and y2, then letting w = y1− y2, it follows w′+a(t)w = 0 and w(r) = 0. Then multiplying both sides of the differential equation by eA(t)

where A′ (t) = a(t) , it follows(

eA(t)w)′

= 0 and so eA(t)w(t) = C for some constant, C.

However, w(r) = 0 and so this constant can only be 0. Hence w = 0 and so y1 = y2. ■Finally, consider the general linear initial value problem.

Definition 29.1.9 A linear differential equation is one which is of the form

y′+a(t)y = b(t)

where a,b are continuous. The corresponding initial value problem is

y′+a(t)y = b(t) , y(t0) = y0.

Now here are the steps for solving the initial value problem.

1. Find the integrating factor∫

a(t)dt ≡A(t)+C. The integrating factor is exp(A(t))=eA(t).

2. Multiply both sides by the integrating factor.

exp(A(t))(y′ (t)+a(t)y(t)

)=

ddt

(exp(A(t))y(t)) = exp(A(t))b(t)

Why is this so? It involves the chain rule and the product rule.

ddt

(exp(A(t))y(t)) = exp(A(t))A′ (t)y(t)+ exp(A(t))y′ (t)

= exp(A(t))a(t)y(t)+ exp(A(t))y′ (t)

= exp(A(t))(y′ (t)+a(t)y(t)

)

540 CHAPTER 29. FIRST ORDER SCALAR ODE2. Multiply by exp (A (t))chain rule and product ruleexp(A(t)) (0! ba(t)y) = © (exp(A(t))y) =exp(A(s)) (0)3. Do f to both sides. Pick F (t) € fexp(A (t)) b(t) dt.exp(A(t)) y(t) =F ()+Cy(t) =exp(—A(1)) F (t) + Cexp(—A (t))This proves the following theorem.Theorem 29.1.7 The solutions to the equation, y' +-a(t)y = b(t) consist of all functionsof the form y(t) = e AMF (t) +e4MC where F(t) € [eA b(t) dt and C is a constant,A’ (t) =a(t).Finally, here is a uniqueness theorem.Theorem 29.1.8 If a(t) is a continuous function, there is at most one solution to the initialvalue problem, y' +a(t)y = b(t), y(r) =yo.Proof: If there were two solutions y; and yz, then letting w = yj —y2, it follows w! +a(t)w =O and w(r) = 0. Then multiplying both sides of the differential equation by e4/where A’ (t) = a(t), it follows (ew) = 0 and so e4()w(t) = C for some constant, C.However, w(r) = 0 and so this constant can only be 0. Hence w = 0 and so y; = yo.Finally, consider the general linear initial value problem.Definition 29.1.9 A linear differential equation is one which is of the formy'+a(t)y=d(t)where a,b are continuous. The corresponding initial value problem isy'+a(t)y=b(t), y(t) =yo.Now here are the steps for solving the initial value problem.1. Find the integrating factor [ a(t) dt =A (t) +C. The integrating factor is exp (A (t)) =AM,2. Multiply both sides by the integrating factor.cs (exp (A (t)) y(t) = exp (A (t)) b(t)exp (A(t) (v() +a()y() = 5Why is this so? It involves the chain rule and the product rule.¢ (exp(A())y(t)) = exp(A(r))A’(@)y(1) +exp(A (0) yO)dt= exp(A(t))a(t)y oe (t))¥' (¢)= exp(A(t)) (y'(4) +a(t)y(1))