I analyzed some example of linear equation systems and tried to distinguish the difference between the solve and fsolve in these cases.
1. In general case, the deviation between solve and fsolve is less than 10^-8, for example:
But, in some case, the deviation is more than 10^-7, for example:
how to reduce the error of solving the problem by fsolve?
2. It seems that solve will automatically translate the answer into numerical result while the system is composed with decimals, for example, we redo the second case with no decimals:
Is it possible to get the exact solution while the decimal is involved in the system?
The related file is attached. I'd appreciate any help on this topic. Thank a lot.