2
$\begingroup$

I am new user of Mathematica, admire the software, but encounter the error while trying to solve large systems of equations. For example

A1 = (a/1417 + b*1417 + c)*Exp[d/(e - 190)] == 15.838
B1 = (a/2046 + b*2046 + c)*Exp[d/(e - 190)] == 10.876
C1 = (a/4035.1 + b*4035.1 + c)*Exp[d/(e - 190)] == 6.1272
D1 = (a/6032.4 + b*6032.4 + c)*Exp[d/(e - 190)] == 4.9133
E1 = (a/8043.2 + b*8043.2 + c)*Exp[d/(e - 190)] == 3.9884
FindRoot[{A1, B1, C1, D1, E1}, {{a, 41}, {b, 1}, {c, 23}, {d, 324}, {e, 11}}]

Every time it gives this error:

FindRoot::jsing: Encountered a singular Jacobian at the point {a,b,c,d,e} = {41.,1.,23.,324.,11.}. Try perturbing the initial point(s).

My guessed values are close to the solution, which I got from matlab. Can anyone help me with this issue?

$\endgroup$
  • 4
    $\begingroup$ since the quantity Exp[d/(e - 190)] is the same in every equation you effectively have only 4 unknowns. $\endgroup$ – george2079 Feb 8 '18 at 19:58
2
$\begingroup$

A naive approach, solving the equations one by one:

A1 = (a/1417 + b*1417 + c)*Exp[d/(e - 190)] - 15.838
B1 = (a/2046 + b*2046 + c)*Exp[d/(e - 190)] - 10.876
C1 = (a/4035.1 + b*4035.1 + c)*Exp[d/(e - 190)] - 6.1272
D1 = (a/6032.4 + b*6032.4 + c)*Exp[d/(e - 190)] - 4.9133
E1 = (a/8043.2 + b*8043.2 + c)*Exp[d/(e - 190)] - 3.9884

a = a /. First@Solve[A1 == 0, a]
b = b /. First@Solve[B1 == 0, b]
c = c /. First@Solve[C1 == 0, c]
Simplify[D1]
(* 0.375437 *)

So there is no chance that D1 == 0: your system has no solutions.

By the way, the guessed values are not "close to the solution", e.g. A1 /. {a -> 41, b -> 1, c -> 23, d -> 324, e -> 11} returns 219.

Edit As per your comment, if you want to minimize the sum of the squares, then do:

 NMinimize[A1^2 + B1^2 + C1^2 + D1^2 + E1^2, {a, b, c, d, e}]
 {0.12353, {a -> 22002.5, b -> 0.000180488, c -> -0.0569829, 
               d -> -4.81166, e -> -11766.7}}

That confirms that there is no root (the cost function would have been 0) but gives you the values that make the equation the "least wrong" (in a certain meaning).

$\endgroup$
  • $\begingroup$ Do you mean no solutions at all? I meant to enter these ones as guessed values: FindRoot[{A1, B1, C1, D1, E1}, {{a, 20617}, {b, 0}, {c, 0}, {d, -26}, {e, 11}}] $\endgroup$ – vladib Feb 8 '18 at 20:35
  • $\begingroup$ @vladib Yes, not solutions at all (except maybe if 0.3757437=0 :)). Your values do not yield zero in {A1, B1, C1, D1, E1} /. {a -> 20617, b -> 0, c -> 0, d -> -26, e -> 11}. Maybe you want to weaken your problem and look for a mimimum instead of a root? $\endgroup$ – anderstood Feb 8 '18 at 20:39
  • $\begingroup$ thanks a lot man, i'll adjust it. $\endgroup$ – vladib Feb 8 '18 at 20:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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