CS 470 Fall 2008
Final Exam Review
The final is comprehensive
You should bring a calculator, and you can bring one 8.5"x11" sheet of notes.
Know the stuff from the previous two review sheets. New stuff is
as follows:
Sequential Choice for Nondeterminstic Worlds
- Markov decision processes: transition probability, discount factor, rewards
- Nondeterministic consequences of choice
- Instantaneous rewards
- Next state
- Definitions and terms
- Recursion relations
- Value iteration
- algorithm
- philosophy
- parameters
- Policy iteration
- algorithm
- philosophy
- parameters
- Q-learning
- weighted averages
- algorithm
- definition of "convergence"
- conditions for convergence
Multi-Agent Choice
- Framework: S, A1, A2, C, G1, G2,
U1, U2,etc.
- Payoff matrices: non-zero sum games, zero-sum games, general-sum
games
- Extensive-form games (turn taking)
- Minimax
- Value
- Solution
- Alpha-Beta pruning: know algorithm and be able to apply it
- Evaluation functions
- Definitions of a solution
- Minimax
- Nash equilibrium
- Strategic dominance
- Best response
- Pareto dominance