Solving coupled ODE symbolically for the parameters

I'm trying to solve the following system:

$$\dot{x_1} = -i w_1 x_1+ i s_1 x_2$$ $$\dot{x_2} = -i w_2 x_2- i s_2 x_1$$

For $$X(t)_{1,2}$$, where $$x_{1,2} = X(t)_{1,2}\cdot e^{ip(t)_{1,2}}$$

I know the solution is:

$$\dot{X_1}=s_1X_2sin(p_1 - p_2)$$ $$\dot{X_2}=s_2X_1sin(p_1 - p_2)$$

I tried the following:

x1 = X1[t]*E^(I*p1[t]);
x2 = X2[t]*E^(I*p2[t]);

x1dot = -I*w1*x1 + I*s1*x2;
x2dot = -I*w2*x2 + I*s2*x1;

x1d = D[x1, t];
x2d  = D[x2, t];

{{X1s[t], X2s[t]}} =
{X1[t], X2[t]} /.
Simplify[
Solve[
{x1d == x1dot, x2d == x2dot},
{X1[t], X2[t]}]]


But The result I get is:

$$-\frac{i \left(\text{x1d} \text{p2}'(t)+\text{s1} e^{i \text{p2}(t)} \text{X2}'(t)+\text{w2} \text{x1d}\right)}{e^{i \text{p1}(t)} \left(-\text{w1} \text{p2}'(t)+\text{s1} \text{s2}-\text{w1} \text{w2}\right)},-\frac{i \left(\text{w1} \text{X2}'(t)+\frac{\text{s2} \text{x1d}}{e^{i \text{p2}(t)}}\right)}{-\text{w1} \text{p2}'(t)+\text{s1} \text{s2}-\text{w1} \text{w2}}$$

I'm not sure if Mathematica can use Euler's formula to simplify the complex exponents and get to the known solution. I guess it's possible but I'm not sure how to do it

One way to derive the "known solution" is to substitute the expressions for $$X_{1,2}(t)$$ into the differential equations for $$x_{1,2}$$ and solve for the derivatives $$\dot{X}_{1,2}$$. Then take the real parts of $$\dot{X}_{1,2}$$, like this:

ClearAll["Global*"]

x1 = X1[t] Exp[I p1 t];
x2 = X2[t] Exp[I p2 t];

eqns = {D[x1, t] == -I ω1 x1 + I s1 x2,
D[x2, t] == -I ω2 x2 - I s2 x1};

soln = First@Solve[eqns, {X1'[t], X2'[t]}];

X1'[t] /. soln // Re // ComplexExpand // Simplify (* s1 Sin[(p1 - p2) t] X2[t] *)
X2'[t] /. soln // Re // ComplexExpand // Simplify (* s2 Sin[(p1 - p2) t] X1[t] *)


Of course, you would need a good argument for taking only the real parts.

• Thank you! Can you please explain what First@Solve is doing? May 26 at 11:52
• The First@Solve can, probably, be understood most easily by running the code with the semicolons removed and again with the First@ part removed and comparing the results. The Solve part returns a list of solutions for the derivatives $\dot{X}_{1,2}$ and the First@ takes only the first solution. In this case there is only one solution, so the effect of First@ is similar to using Flatten`. Thanks for accepting this answer. May 26 at 20:12