Lab 17: Utility
- Write a program that calculates the utility of
a bet that paid off 22% of the time 75£.
- Now replace this with user input for one percentage and
one utility value.
- Now write a system that predicts the value of a 2-armed
bandit problem.
- Build the probabilities and values in. (Start with a constant
value output. E.g 10% 50£ vs. 90% 5£)
- Run the system and keep track of the results.
- Can your system pick the optimal move?
- What happens if the values change?
- If your system is trying to optimise expected value, does
it ever pick a suboptimal outcome?