What is multicriteria decision analysis?

MCDA is a technique aimed for the decision makers who are faced with numerous and conflicting evaluation to take a decision.

We are doing lot of multi-criteria decision making in our day to day life.  For instance if you want to buy a house, you will have ‘m’ number of criteria’s and at the same time there will be ‘n’ number of stakeholder who influence directly or indirectly to your decision. Stakeholders are the people who involved in the decision making process. This include you, your wife, father, mother, renting people you are giving your house for rent, your employer etc.

I will give a vague idea how to do MCDA? The first process to do is do a “C.A.U.S.E” analysis. Criteria, Alternatives, Uncertainty, Stakeholder and Environmental / External entity.  List down all the factors for CAUSE, bring down the stakeholders and prioritize them considering their power and interests. After CAUSE analysis you have to list down and form all clusters.

The second process is constructing a decision tree, see the decision tree below for the example mentioned above. Once you build decision tree cross check with the cluster if there is any point missed out.

The next step is to assign values for the alternatives for the leaf nodes from values 0 to 100. For most feasible one you should give 0 and for least feasible one you should give 1oo. When we assign values to the decision tree take consider things in mind, the values given should be in the ratio of the cost affected / criteria affects you or the decision.

The fourth step is to give scores or weight to the nodes from right to left. This can be done by considering all the stakeholders and their scores to the nodes. You can take the average of the weights of all stakeholders. If there are numerous nodes in the right hand side  select few which is of top priority and assign the weights. After that you can assign values to other by considering and assigning values to its families.

Once weightage is done you will get a alternatives in the top position according to the weight given. The next step is to analyze robustness of the decision. This can be done by analyzing the sensitivity across different criterion. If majority of the sensitivity graphs shows the same result as final decision you can say that your decision is robust.

Lots of tools are available to do multi-criterion decision analysis. I am using the software called VISA.