
Ryan C. answered 01/18/23
Ivy League Professor | 10+ Years Experience | Patient & Kind
Let's break this problem into steps.
Step 1. Separate your variables.
Let u(x,t) = X(x)T(t). Plug this into your PDE and separate the X and T dependencies. You'll eventually be left with
T'(t)/(4*T(t)) = X''(x)/X(x).
I grouped the coefficient 4 with the T dependence for convenience. We introduce the separation constant s.
T'(t)/(4*T(t)) = X''(x)/X(x) = s.
This gives us two ODE's X''(x) = s*X(x) and T'(t) = 4*s*T(t).
Step 2. Solve the X(x) equation with the appropriate boundary conditions.
It is not hard to show that the boundary conditions u(0,t) = 0 and u(1,t) = 0 imply X(0) = 0 and X(1) = 0, respectively. Thus, we need to find the solution of X''(x) = s*X(x) that satisfies the boundary conditions X(0) = X(1) = 0.
I'll leave it to you to show that, when s >= 0, the only solution to this boundary value problem is X(x) = 0 for all x. This is a trivial solution that we throw out. When s < 0, we get nontrivial solutions. Indeed, when s < 0,
X(x) = C1*cos(sqrt(-s)*x) + C2*sin(sqrt(-s)*x)
is the general solution of X''(x) = s*X(x), where C1 and C2 are arbitrary constants. If X(0) = 0, then hopefully you can see C1 = 0. If X(1) = 0, then we have C2sin(sqrt(-s)) = 0. We could let C2 = 0, but that would lead us to the trivial solution X(x) = 0 again. Instead, we require sin(sqrt(-s)) = 0. This implies sqrt(-s) = pi*n, where n is any natural number. Thus, s = -(pi*n)^2. Thus, we have infinitely many nontrivial solutions to our boundary value problem:
Xn(x) = Cn*sin(pi*n*x),
where Cn is arbitrary for each n.
Step 3. Obtain the corresponding time dependence T(t) for the solutions of the boundary value problem.
We've seen that Xn(x) = Cn*sin(pi*n*x) are the nontrivial solutions of the boundary value problem. Their corresponding time dependence solves the ODE T'(t) = 4*s*T(t). The general solution of this ODE is
T(t) = D*exp(4*s*t),
where D is an arbitrary constant. For the time dependence of our nontrivial boundary value problem solutions, s = -(n*pi)^2. Thus, the corresponding time dependence of the boundary value problem solutions is
Tn(t) = Dn*exp(-4*(n*pi)^2*t),
where Dn is an arbitrary constant for each n.
Step 4. Superimpose your nontrivial solutions of the PDE.
In Steps 2-3, we found Xn(x) = Cn*sin(pi*n*x) and Tn(t) = Dn*exp(-4*(n*pi)^2*t). Thus, we have the following nontrivial solutions of the PDE:
un(x,t) = Xn(x)Tn(t) = Cn*sin(pi*n*x) Dn*exp(-4*(n*pi)^2*t).
Since Cn and Dn are arbitrary, we can fold Dn into Cn so that
un(x,t) = Cn*sin(pi*n*x)*exp(-4*(n*pi)^2*t).
We superimpose these solutions to get a new solution
u(x,t) = ∑ un(x,t) = ∑Cn*sin(pi*n*x)*exp(-4*(n*pi)^2*t).
Here, the sum goes from n = 1 to n = ∞.
Step 5. Use Fourier series and the initial condition to determine the arbitrary constants in your solution u(x,t).
The initial condition enforces
u(x,0) = x(1-x) = ∑Cn*sin(pi*n*x).
Looking carefully at the second equality, we see that the Cn are simply the Fourier sine series coefficients of the function x(1-x) defined on the interval [0,1]. There is a formula for these coefficients given by
Cn = 2/L*∫0Lx*(1-x)*sin(n*pi*x/L) dx.
Here, L is the length of the interval, i.e., L = 1. One can evaluate this integral above by repeated integration by parts. One finds
Cn = (4(-1)n+1+4)/(pi*n)3.
Thus, the final solution of this problem is
u(x,t) = ∑(4(-1)n+1+4)/(pi*n)3*sin(pi*n*x)*exp(-4*(n*pi)^2*t).

Ryan C.
It's at the bottom.01/26/23
Aznan M.
What is the answer? please help me sir. thank you01/26/23