Simplifying a symbolic matrix-inverse

Question

I have an $8\times8$ symbolic matrix, $G$. It contains mostly real parameters but a few complex ones too. Mathematica computes (actually pretty fast) its inverse $G^{-1}$. The problem is that the inverse provided is too complicated, that is, it is far from being simplified, even though I am pretty sure that it could be much simpler.

If I apply directly Simplify[]of FullSimplify[], Mathematica gets stuck.

What can I do to massage $G^{-1}$ so that a simplification can be carried on?

I tried basically two things: (I)Inputting some assumptions in the simplification and (II)specifying different Methods-->" " for Inverse[] . Using my assumptions about parameters (positive, reals, etc) leads equally to a stuck calculation and different methods lead to equally complicated inverses.

Another question that may help me solving the issue is:

How can I trace which manipulations are slowing down (Full)Simplify ?

Here is my matrix G, in input form:

{{-γa - 2*ΓHa*Re[tH] - 2*ΓVa*Re[tV], -1/2*γb - ΓHb*Re[tH] - ΓVb*Re[tV], 0, 0, -(tH*Sqrt[ΓHa]*Sqrt[ΓHb]) - tV*Sqrt[ΓVa]*Sqrt[ΓVb] - (Sqrt[ΓHa]*Sqrt[ΓHb]*Conjugate[tH])/2 - 
   (Sqrt[ΓVa]*Sqrt[ΓVb]*Conjugate[tV])/2, 0, 0, -1/2*(tH*Sqrt[ΓHa]*Sqrt[ΓHb]) - (tV*Sqrt[ΓVa]*Sqrt[ΓVb])/2 - Sqrt[ΓHa]*Sqrt[ΓHb]*Conjugate[tH] - 
   Sqrt[ΓVa]*Sqrt[ΓVb]*Conjugate[tV]}, {-1/2*γa - ΓHa*Re[tH] - ΓVa*Re[tV], -γb - 2*ΓHb*Re[tH] - 2*ΓVb*Re[tV], 0, 0, 
  -1/2*(tH*Sqrt[ΓHa]*Sqrt[ΓHb]) - (tV*Sqrt[ΓVa]*Sqrt[ΓVb])/2 - Sqrt[ΓHa]*Sqrt[ΓHb]*Conjugate[tH] - Sqrt[ΓVa]*Sqrt[ΓVb]*Conjugate[tV], 0, 0, 
  -(tH*Sqrt[ΓHa]*Sqrt[ΓHb]) - tV*Sqrt[ΓVa]*Sqrt[ΓVb] - (Sqrt[ΓHa]*Sqrt[ΓHb]*Conjugate[tH])/2 - (Sqrt[ΓVa]*Sqrt[ΓVb]*Conjugate[tV])/2}, 
 {0, 0, -1/2*γa - tH*ΓHa - tV*ΓVa - I*Δa - (I/2)*Δb, -(tH*Sqrt[ΓHa]*Sqrt[ΓHb]) - tV*Sqrt[ΓVa]*Sqrt[ΓVb], 0, 0, 0, 0}, 
 {0, 0, -(tH*Sqrt[ΓHa]*Sqrt[ΓHb]) - tV*Sqrt[ΓVa]*Sqrt[ΓVb], -1/2*γb - tH*ΓHb - tV*ΓVb - (I/2)*Δa - I*Δb, 0, 0, 0, 0}, 
 {-(tH*Sqrt[ΓHa]*Sqrt[ΓHb]) - tV*Sqrt[ΓVa]*Sqrt[ΓVb], -(Sqrt[ΓHa]*Sqrt[ΓHb]*Conjugate[tH]) - Sqrt[ΓVa]*Sqrt[ΓVb]*Conjugate[tV], 0, 0, 
  -1/2*γa - γb/2 - tH*ΓHb - tV*ΓVb - (I/2)*Δa + (I/2)*Δb - ΓHa*Conjugate[tH] - ΓVa*Conjugate[tV], 0, 0, 0}, 
 {0, 0, 0, 0, 0, -1/2*γa + I*Δa + (I/2)*Δb - ΓHa*Conjugate[tH] - ΓVa*Conjugate[tV], -(Sqrt[ΓHa]*Sqrt[ΓHb]*Conjugate[tH]) - Sqrt[ΓVa]*Sqrt[ΓVb]*Conjugate[tV], 0}, 
 {0, 0, 0, 0, 0, 0, -1/2*γb + (I/2)*Δa + I*Δb - ΓHb*Conjugate[tH] - ΓVb*Conjugate[tV], 0}, {-(Sqrt[ΓHa]*Sqrt[ΓHb]*Conjugate[tH]) - Sqrt[ΓVa]*Sqrt[ΓVb]*Conjugate[tV], 
  -(tH*Sqrt[ΓHa]*Sqrt[ΓHb]) - tV*Sqrt[ΓVa]*Sqrt[ΓVb], 0, 0, 0, 0, 0, -1/2*γa - γb/2 - tH*ΓHa - tV*ΓVa - (I/2)*Δa + (I/2)*Δb - ΓHb*Conjugate[tH] - ΓVb*Conjugate[tV]}} 

Where $\gamma a,~\gamma b,~\Gamma Ha,~\Gamma Hb,~\Gamma Va,~\Gamma Vb,~\Delta a,~\Delta_b >0$ and $tH,~tV \in \mathbb{C}$

Answer 1

Create an image-array of your matrix, generate a rule for substition of created variables, invert this image-array, simplify it and reinsert original expressions. (I renamed your matrix G to mat and substituted the two complex variables by their real and imaginary part.)

vars = {tH1, tH2, tV1, tV2, \[Gamma]a, \[Gamma]b, \[CapitalGamma]Ha, \[CapitalGamma]Hb, \[CapitalGamma]Va, \[CapitalGamma]Vb, 
\[CapitalDelta]a, \[CapitalDelta]b}; 

MatrixForm[
mat2 = (Simplify[#1, Thread[vars \[Element] Reals]] &)[
mat /. {tH -> tH1 + I tH2, tV -> tV1 + I tV2}]];

arr = Array[a, {8, 8}];

(arr2 = mat2 arr /. _*a[u_, v_] -> a[u, v]) // MatrixForm

varsarr2 = Variables[Flatten[arr2]]

rule = Flatten@MapThread[Rule, {arr, mat2}, 2]

invarr2 = Inverse[arr2]

Denominator for all elements is same. Finaly show all elements, that are not zero.

den = invarr2[[1, 1]] // Denominator;

(invarr3 = invarr2 /. 1/den -> 1/dd // Simplify) // MatrixForm

(invmat = (invarr3 /. dd -> den // Simplify) /. rule // 
Simplify[#, Thread[vars \[Element] Reals]] &) // Timing   (*   16 sec   *)

Manipulate[invmat[[Sequence @@ k]], {k, List @@@ varsarr2}, 
   ControlType -> Setter]