Thank you for your help
I have another question:
why it improves the speed of calculation after substituting x/p by rho?
Is there any general experience to improve calculation speed?
Thank you very much

Thank you for your help
I have another question:
why it improves the speed of calculation after substituting x/p by rho?
Is there any general experience to improve calculation speed?
Thank you very much

@Preben Alsholm Thank you for your reply

@Preben Alsholm Thank you for your reply

Thank you for your help.

Gantmacher's method means it expands ln(1+x) at point x=2,implicitly (the form is ln(3)+1/3*(x-2)-(1/18)*(x-2)^2+(1/81)*(x-2)^3+...) .So the convergent domain is (1,3) which just convers x=2.

How about a matrix G := Matrix(3, 3, {(1, 1) = 218/29, (1, 2) = 96/29, (1, 3) = 0, (2, 1) = 120/29, (2, 2) = 130/29, (2, 3) = 0, (3, 1) = 0, (3, 2) = 0, (3, 3) = 5}) whose eigenvalues are 2,5 and 10. They can't be convered by any single domain correspanding to any point at which you expand. The command MatrixFunction and the Gantmacher's method seem to give different answers.

PS: In maple 15 I can't find Gantmacher in help page. And it is night in my half sphere. I have to go to sleep. See you tomorrow.

Thank you for your help.

Gantmacher's method means it expands ln(1+x) at point x=2,implicitly (the form is ln(3)+1/3*(x-2)-(1/18)*(x-2)^2+(1/81)*(x-2)^3+...) .So the convergent domain is (1,3) which just convers x=2.

How about a matrix G := Matrix(3, 3, {(1, 1) = 218/29, (1, 2) = 96/29, (1, 3) = 0, (2, 1) = 120/29, (2, 2) = 130/29, (2, 3) = 0, (3, 1) = 0, (3, 2) = 0, (3, 3) = 5}) whose eigenvalues are 2,5 and 10. They can't be convered by any single domain correspanding to any point at which you expand. The command MatrixFunction and the Gantmacher's method seem to give different answers.

PS: In maple 15 I can't find Gantmacher in help page. And it is night in my half sphere. I have to go to sleep. See you tomorrow.