- #1
TheDestroyer
- 402
- 1
Hello guys,
I'm writing a C++ function for matrix inversion. I saw many algorithms online, but I'm concerned about the case when a diagonal element equals zero, which is, for some reason, not taken into account by those websites.
When we do the elimination for solving a system of linear equations, it's fine to exchange 2 lines, since the columns represent the variables for which the equations are solved.
But in the case of inverting the matrix to multiply it with another column matrix, which is a vector whose rows have elements that comply to the columns in that matrix I want to invert (I need this for Levenberg-Marquardt algorithm). How can I solve this problem? can I simply do the reciprocal exchange after the inverting is done (e.g., start exchange 1 with 2, and at the end redo the exchange to restore it the way it was)? could someone please explain the theoretical scheme of the problem? (I mean when the matrix is invertible, what kind of conditions will I get in the matrix elements).
Thank you for any efforts :)
I'm writing a C++ function for matrix inversion. I saw many algorithms online, but I'm concerned about the case when a diagonal element equals zero, which is, for some reason, not taken into account by those websites.
When we do the elimination for solving a system of linear equations, it's fine to exchange 2 lines, since the columns represent the variables for which the equations are solved.
But in the case of inverting the matrix to multiply it with another column matrix, which is a vector whose rows have elements that comply to the columns in that matrix I want to invert (I need this for Levenberg-Marquardt algorithm). How can I solve this problem? can I simply do the reciprocal exchange after the inverting is done (e.g., start exchange 1 with 2, and at the end redo the exchange to restore it the way it was)? could someone please explain the theoretical scheme of the problem? (I mean when the matrix is invertible, what kind of conditions will I get in the matrix elements).
Thank you for any efforts :)