I have an equation F(x,y;a,b,c,…)=0, where x and y are variables, a,b,c,… - parameters. y is a real number, x - complex. For every given y I need to solve this equation, i.e. to find Im(x) and Re(x) (this equation may have many solutions). After solving this equation I need to plot two graphs: the dependense of y from Im(x) and y from Re(x).
For example, F(x,y)=x^2 + sin(xy).
If y and x were real numbers, it would be possible to use ContourPlot. However, as my variables are complex, I have to solve this equation numerically and after that plot my graphs.
I have never come across such a problem (and also to date I have a little experience in Mathematica), so could anyone tell me how to solve this problem?
I need the solution of the equation F[x,y]=0, i.e. Re(F[x,y])=0 and Im(F[x,y])=0 simultaneously. I mean, for example, I take y_1=0 and find x_1 = Im[x_1]+iRe[x_1] (maybe more solutions x_1’, x_1’‘,…), then I take y_2=0.1 and find x_2 = Im[x_2]+iRe[x_2] (maybe also x_2’,…) and so on. After that (using these numbers) I plot the dependense of y_i on Im[x_i] (also on Im[x_i’],…) and on Re[x_i] (again on Re[x_i’],…).
Regarding specific example, I consider this equation Tanh[Sqrt[-x^2 - y^2]0.3](y^2 + igy - 4)Sqrt[-x^2 - y^2] + (y^2 + igy)Sqrt[-x^2 - y^2 ((y^2 + igy - 4)/(y^2 + igy))]] == 0, where {g,0,0.2}, i is a complex unit, Tanh is a hyperbolic tangent. I would like to take, for example, g=0.1 and for y from 0 to 6 find x=Re[x]+i*Im[x]. After I find these solutions I would like to plot the dependense of y on Re[x] and the dependense of y on Im[x].
Dear Daniel, thank you so much for your help! Hovewer, I am not sure that this is exactly what I need, because I don’t quietly understand how it works.
I can explain what I need in other words. I need the graph of the equation F[x,y]=0, x is a complex number, y is a real number, {y,0,8}. The graph of this equation is 3D, because we have 3 variables Re[x], Im[x], y. I need two 2D graphs y as a function of Re[x] and Im[x], I mean that I need two projection of this 3D graph.
Thank you. But can I further extract two 2D graphs from this 3D graph? I mean I can plot F[Re[x],Im[x],y]=0, but I would like to get two projections of this graph: Re[x],y and Im[x],y.
You might do better to define a root-finding function that, given a value for y, finds Re[x] and Im[x]. Using the redefined form I showed earlier, with s and t as variables for the real and imaginary parts of x respectively, this might be done as below. I use random starting points so as to have a chance at sampling from different branches of the solution curve.
Dear Daniel, could you clarify how do you do this? I try to use your code, but I don’t get a result.g = 1/10; expr = x - Sqrt[y^2 + I*y]; e2 = expr /. x -> s + I*t; {re, im} = ComplexExpand[{Re[e2], Im[e2]}]; root[yval_?NumericQ] := {s, t} /. FindRoot[{re == 0, im == 0} /. y -> yval, {s, RandomReal[{-5, 5}]}, {t, RandomReal[{-5, 5}]}] pts = Quiet[Table[root[j], {j, .01, 8, .01}]]; ListPlot[pts]
Dear Daniel, I am sorry for interruption. I don’t know what was the problem, but now everything is fine. Could you answer a couple of other questions? Firstly, what shoul I do to consider this problem only in the region {Im[x],0,8}, {Re[x],0,8}, {y,0,8}. And how can I distinguish between the dependense of y on Re[x] and y on Im[x] depicted on the figure? Also, could you clarify what do I change when alter RandomReal[{0, 5}] and {j, .01, 5, .01}?
Dear Daniel. I have decided to check your method I have considered a very simple problem (x=f[y]) which can be easily solved by a very straightforward method (just using Plot). I don’t know why but the results are essentially different.
I attach my file, but the general idea is very simple. If we have x = f[y] and x=x’+ix’‘, we can just plot x’=Re[f[y]] and x’'=Im[f[y]] and compare these graphs with the graphs which we get using your method.
The plots look more similar if you recognize that the equivalent form for the second one should use ParametricPlot instead of Plot (such distinctions are very important). The ListPlot will need a range setting as well. Try the two below.
No, no, no. Dear Daniel, I am sorry that my explanations were probably unclear. I need exactly what I have shown you in my version (because I know the answer for this particular example). Now I attach a picture where I explain everything in symbols (and in a couple of words). I think that this picture makes everything much more clear without words (I am sorry that I have presented in just on paper, but I think it is clear enough).
However, it would good to try to explain again what I need by words. I have a function f[x,y], where y is real and x is complex. I need to solve this equation for y from 0 to 10, i.e. to find x for every y from 0 to 10. After getting the solutions (Re(x), Im(x), y)I need to plot two graphs: y as a function of Re(x) and y as a function of Im(x) (i.e. use the numbers from 1st and 3d column for the first graph and the numbers from 2nd and 3d column for the second graph)
I was simply explaining why the plots were different.
As for getting y as functions, respectively, of Re[x] and Im[x], that gets tricky because it is in general not a function but rather a “multivalued function”. You will probably want to get a continuous branch of such inverses. One method you might consider is to find a value for y at a given re(x), say, and then trace from there by setting up an NDSolve equation.
Lets define the real and imaginary parts of f as ref and imf respectively. And the real and imaginary parts of x we will call rex and imx. The equations you have are {ref[rex,imx,y]==0,imf[rex,imx,y]==0}. To inverst y in terms of rex, treat it as a function of rex (the whole point of the exercise) and differentiate. I’ll show explicit code for your example.
