My problem is fitting two complex-valued functions f and g:
|f(x,a,b,c)-g(x)|^2 < epsilon 
I want to find best approximation for a, b and c where parameters are real.
Can you suggest the method to solve this problem?
I have the analytic form of the functions f and g.

I want to find the first positive solution of the system of trigonometric equations inside the loop.

The solutions are in the form of "d=number*_Z +number" but I need one exact solution to use it for next run of the loop.
 

restart;
L[0]:=0:
for i from 1 by 1 to 3 do
assume(0<d[i], d[i]<1):
assume(-0.01<a[i], a[i]<0):
L[i]:= L[i-1]+ d[i]:
sys[i]:={Re((-80*Pi*I*a[i]/((a[i]+1)^3))*exp(4*Pi*I*L[i])) = -0.4, Im((-80*Pi*I*a[i]/((a[i]+1)^3))*exp(4*Pi*I*L[i])) = 0.8}:
solve(sys[i], {a[i],d[i]}, useassumptions = true,AllSolutions=true):
end do;
 

These are the solutions:

d[1] = 0.03689590440 + 0.5000000000 _Z1

d[2] = -1.000000000 d[1] + 0.03689590440 + 0.5000000000 _Z2

d[3] = -1.000000000 d[1] - 1.000000000 d[2] + 0.03689590440 + 0.5000000000 _Z3

I want to solve this equation with assumptions!!!

restart;
assume(d::real, d>0):
assume(a::real, -0.01 < a, a < 0):
sys:={-800*Pi*a*cos(6.557377048*Pi*(3.470797713+d))/(a+1)^3 = -.9396060697, 800*Pi*a*sin(6.557377048*Pi*(3.470797713+d))/(a+1)^3 = -.3238482794};
solve(sys, {a,d},useassumptions=true, AllSolutions=true);

one of the solutions has true "a" but "d" is wrong, I want one true solution!