One of the tasks in my thesis is make a decision matrix. I have all the criteria from C1 to Cn (n number of criteria) and made a pairwise matrix. Here's the problem. I can't rely on one person to get the numbers from. I have to rely on 5 people. Let's consider the Importance of Criteria 1 to Criteria 2. What If Person 1 says It's 7 times more important but P2 says it's 5 times, P3 says it's 6 times, P4 says and so on. All of them have different answers. I don't know what to do if that happens.
Dear Ramon Araneta,
I suggest you to see links and attached files on topic.
https://backend.orbit.dtu.dk/ws/portalfiles/portal/104276012/DTU_Transport_Compendium_Part_2_MCDA_.pdf
https://www.win.tue.nl/eis2010/lib/exe/fetch.php?media=2009:z-ssebuggwawo.pdf
Article Analytic hierarchy process application in selecting the mode...
Article Group Aggregation Techniques for Analytic Hierarchy Process ...
Best regards
More decision-makers (DMs) will help to reduce the vagueness in decision-making. One can use the group decision making employing AHP. The Geometric Mean (GM) method may be used. Please refer to research on Evaluating critical success factors in implementing the E-learning system using multi-criteria decision-making .
Hi Mr Areneta
My suggestion is:
Reach an average of expert opinions with one of the available methods.
Most of the time, I take the geometric mean from the opinions of a large number of experts and reach for a final matrix, and finally, with this final matrix, I go for decision-making and fuzzyness.
Good luck
I would look up group decision making approaches where decision makers have different weights in the process. This one could help
Duman, G. M., El-Sayed, A., Kongar, E., & Gupta, S. M. (2019). An intelligent multiattribute group decision-making approach with preference elicitation for performance evaluation. IEEE Transactions on Engineering Management.
Good morning, Mr. Araneta
This is a problem of inconsistent opinions, something that seems to be common when applying peer analysis. In order to avoid this situation, I like to use the Expert Choice software trial version, as it shows, at the moment the participant is responding, if there is inconsistency in his opinion. Thus, it is easy to correct the incidence the moment it occurs.
Good job!
The fundamental concern is to obtain accurate data for each criteria. Now matter what the source of data, there is always statistics associated with the measurement. The first opportunity is to formulate the criteria so the measurement statistics can be obtained. Once the measurements and associated statistic are produced, the assessment may be carried out to produce a decision candidate. The decision sensitivity may be assessed for each of the criteria. If the statistics for the given criteria indicate high sensitivity such that the reliability of the decision is questionable to the DM, further work on that criteria must be carried out.
When framing a decision, a consideration is how much the decision process will cost. If we spend a lot of resources assuring that each criteria will be have low variance, we may be spending more than necessary to achieve sufficiently accurate results. By establishing the criteria method to low resolution, we may minimize the effort necessary to achieve the result. Then spend the effort to obtain better resolution where the decision process indicates the most value.
This may further be refined by developing criteria that may be more objectively measured.
If we are using SMEs' opinions for our decisions, we are outsourcing our decision process. This is not sustainable. A more sustainable approach is to establish criteria that can be more accurately measured and testable such that statistic can be obtained.
- Badges
- Science topic
- Similar topics
- Mathematics
- Algebra
- Matrix