Discussion & Conclusions
- In terms of performance with respect to processing time, Method A - Matlab's 'backslash' command is the hand's down winner, followed by Method B - Matlab's inverse command. Use of Method C - Cramer's Rule is a very distant third, especially for large matrices (high values of n).
- It was not possible to perfom any Method A, B, or C comparisons for n=10,000; this was due only to insufficient computer memory.
- In terms of performance with respect to accuracy, without comparing every single vector element to its peer in the respective other methods, I used the 2-norm to deduce if the solution vectors were roughly the same. For low values of n (n=10, n=100), the 2-norms were the same for all three methods.
- The 2-norms for Method A and B were equivalent. From this one can infer equal degrees of accuracy.
- The 2-norm computed for a moderately large system (n=1000) "blew up" , returning the value NaN. I cannot provide an explanantion for this anomaly.
- General conclusion for large or small system of equations, Method A rules supreme and is preferred! Method C is deprecated.