Dsolve problem with a system of 3 differential equations

I'm not a mathematician so I apologize if what I'm asking is trivial. I'm trying to solve a system of three differential equations: $$\frac{dA}{d t} = -k_1A(t)+k_{-1}B(t)-nA(t)\frac{dC}{d t}$$ $$\frac{dB}{d t} = k_1A(t)-k_{-1}B(t)-(nA(t)-1)\frac{dC}{d t}$$ $$\frac{dC}{d t} =k_2B(t)A(t)^{nA(t)}$$ with the conditions $$B(0) = 0$$ and $$C(0) = 0$$

However, dsolve does not provide an answer - it just runs for a long time. Any help would be appreciated!

Here's the code used in Mathematica:

DSolve[{a'[t] == -k1 a[t] + k11 b[t] - n a[t] c'[t],
b'[t] == k1 a[t] - k11 b[t] + (n a[t] - 1) c'[t],
c'[t] == k2 b[t] a[t]^(n a[t]), b == 0, c == 0},
{a[t], b[t], c[t]}, t]

More information that might be relevant: k1, k11 and k2 are positive and real, typically between 0 and 1. n is between 1 and 4 or so.

And a follow up question - can I use ndsolve to determine the above mentioned parameters, given some c[t] datapoints?

• Please add the code you used in Mathematica. Because a) we're lazy and b) there might be typos – AndreasP Oct 18 '16 at 14:23
• Here: DSolve[{a'[t] == -k1 a[t] + k11 b[t] - n a[t] c'[t], b'[t] == k1 a[t] - k11 b[t] + (n a[t] - 1) c'[t], c'[t] == k2 b[t] a[t]^(n a[t]), b == 0, c == 0}, {a[t], b[t], c[t]}, t] – Igor Oct 18 '16 at 14:25
• What are the values of k1 etc.? – corey979 Oct 18 '16 at 14:38
• The system is nonlinear, so a closed-form solution may not exist. Try NDSolve to get a numerical solution. – Chris K Oct 18 '16 at 14:41
• Why not immediately eliminate C? $C$ itself doesn't matter to $dA/dt$ and $dB/dt$---only $dC/dt$ does, and you have an expression for that quantity. I still don't think there will be a closed-form solution (because of $A^A$) but Mma should be able to find a numeric solution. – evanb Oct 18 '16 at 15:47

This system can be solved numerically and generally with ParametricNDSolve, if you specify an initial condition for a (which I picked to be 1 below for concreteness).

soln = ParametricNDSolve[{a'[t] == -k1 a[t] + k11 b[t] - n a[t] c'[t],
b'[t] == k1 a[t] - k11 b[t] + (n a[t] - 1) c'[t],
c'[t] == k2 b[t] a[t]^(n a[t]), b == 0, c == 0,
a == 1}, {a[t], b[t], c[t]}, {t, 0, 10}, {k1, k11, k2, n}]

Now you can play with various values of the coefficients

Manipulate[Plot[Evaluate[
Through[({a[t], b[t], c[t]} /. soln)[k1, k11, k2, n]]],
{t, 0, 10}, PlotRange -> All],
{k1, 0, 1}, {k11, 0, 1}, {k2, 0, 1}, {n, 1, 4}]

where I chose the ranges of the parameters based on your suggestion. This gives a little Manipulated plot: To get the functions themselves, you can do eg.

{AA[t_], BB[t_], CC[t_]} := Evaluate[
Through[({a[t], b[t], c[t]} /. soln)[k1, k11, k2, n]]]

where you plug in concrete numerical values of k1, k11, k2, n. Then you can use AA, BB, CC as functions (ie. A will produce a numerical value).