In literature I found the random consistency index (RI) only for 10 or 15 number of criteria. Anyone knows how can I calculate the Consistency Ratio (CR) if I have a matrix nxn, with n greater than 15? Are there any other methods to verify the consistency of the comparison matrix?
Dear Theo
Correct me if I am wrong, but in my opinion for matrices larger than 3 elements the Eigen Value method gives results that do not coincide with the geometric mean
I wonder then why to use the Eigen Vslue method when it is simpler and more accurate to use the geometric mean?
Dear Theo
Thank you for your answer.
Needless to say I plenty agree with you
Your last paragraph is really enlighting and it should be part of the 'MCDM Bible'
It is really hard to imagine that the stability or strength of the best solution, depends on only ONE criterion, when reality shows that there are many criteria, acting simultaneously, with different increments and decrements that should be investigating in a rational sensitivity analysis. This can give the stakeholders much more important information that knowing when the ranking changes, which in addition, is false, on the account that it is based on criteria weights arbitrarily elicited
I honestly doubt that the current procedure followed by all methods in sensitivity analysis, may satisfy stakeholders, and as you hinted, this is paramount.
The actual system is simply a sensitivity analysis of weights, not a sensitivity analysis of the solution, and therefore, irrelevant.
- Badges
- Science topic
- Similar topics
- Indexes