Create a grading system based on Monte Carlo simulation in R to find the best set of weights.
Assume something about the distributions of the grades from 1. exam, 2. assignments, and 3. activity.
Compare different sets of weights.
What will be for you the criterium of the goodness of the results? Maximizing of the expected rating or minimalization of failure?