1
$\begingroup$

I am new to Mathematica (or computations for that matter), can one please tell me how to solve these coupled differential equations, with boundary conditions on infinity, using NDsolve?

x'[t] = x[t]x[t]x[t] -3x[t]y[t]y[t] + x[t] + 1

y'[t] = y[t]y[t] - 3x[t]x[t]y[t] -y[t] - 1

x'[t] represents derivative with respect to t.

with x-> -0.799 as t--> infinity, and y--> -0.304 as t--> infinity.

It would be a big help! Thank you!

$\endgroup$
2
  • 2
    $\begingroup$ First, be sure to use == (not =). And space terms. $\endgroup$ Jul 24, 2021 at 17:46
  • 1
    $\begingroup$ Try sol={x[t],y[t]}/.NDSolve[{x'[t]==x[t]^3-3x[t]y[t]^2+x[t]+1,y'[t]==y[t]^2-3x[t]^2y[t]-y[t]-1,x[101]==-0.799,y[101]==-0.304},{x[t],y[t]},{t,100,101}][[1]];Plot[sol,{t,100,101}] and see if it gives you a plot then very very gently push those 100 and 101 around and see what happens. It looks to me like your system blows up using the default methods. $\endgroup$
    – Bill
    Jul 24, 2021 at 17:47

1 Answer 1

4
$\begingroup$

The problem is ill posed, rhs is not zero at $t\rightarrow\infty$. We can check as follows:

f = x[t] x[t] x[t] - 3 x[t] y[t] y[t] + x[t] + 1;
g = y[t] y[t] - 3 x[t] x[t] y[t] - y[t] - 1;
    
ff = f /. {x[t] -> u, y[t] -> v}
gg = g /. {x[t] -> u, y[t] -> v}
res={u, v} /.NSolve[1 + u + u^3 - 3 u v^2 == 0 && -1 - v - 3 u^2 v + v^2 == 0, {u, v}] // TableForm
 (*
 0.138034               1.65968
-0.317608-0.766191 I    0.524953 +0.508131 I
-0.317608+0.766191 I    0.524953 -0.508131 I
 0.234871 -0.516978 I  -0.769585-0.294646 I
 0.234871 +0.516978 I  -0.769585+0.294646 I
 0.400906 -1.18382 I    0.186183 -0.171329 I
 0.400906 +1.18382 I    0.186183 +0.171329 I
-0.774371              -0.320563
 *)

As you can see, there are no your numbers among the solution.

You can generate the phase portrait of your system with

StreamPlot[{1 + u + u^3 - 3 u v^2, -1 - v - 3 u^2 v + v^2}, {u,-2, 2}, {v, -2, 2}, 
 Epilog -> {PointSize[Large], Red, Point[res[[1]]], Green, Point[res[[8]]]}]

enter image description here

$\endgroup$
2
  • 1
    $\begingroup$ I'd say the last one is probably the branch that the OP is looking for, but had insufficient precision on. $\endgroup$
    – Roman
    Jul 25, 2021 at 10:13
  • 1
    $\begingroup$ @Roman Good problem formulation is the first step to the solution. $\endgroup$
    – yarchik
    Jul 25, 2021 at 10:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.