Use MATLAB

Choose a decision making problem, that could be represented with a decision matrix with at least 4 criteria and 5 alternatives (multiple-criteria decision making under risk, certainty, ignorance,... or a combination of these)

1. Find the ordering of alternatives from the most promising to the least promising one. (this will obviously include finding the weights of criteria somehow, normalizing the evaluations. to criteria, aggregating the evaluations for each alternative, etc.)
2. Remove one of the alternatives (e.g. the best one) and do the same - see, how (and if) the aggregated evaluations of the alternatives change
3. Try to assess how sure you are about the parameters of the model (probabilities, weights of criteria, evaluations. criteria if points are used,...)
4. Find the parameter you are least sure about and suggest how to do sensitivity analysis for this parameter in your model