Essentially, start at the case of one scenario for one iteration and code it. Just be sure to use generalized variable names and functions that can be called for the various scenarios to come.
Then, wrap that in a loop based on the number of cases within a scenario, saving the results in either structures or cell arrays to account for the size differences. You'll have to decide what it is that a "comparison" really means for such cases -- plots are easy; numeric tests may be more of an issue depending on what you're looking at for results.
Once you've got that down, then you can simply add another looping construct over the scenarios. It all sounds more complicated than it likely is given the ease with which Matlab handles reallocation of memory and the difference in types that can be held in cell arrays almost transparently.
