Task Environment PEAS
- Performance measure
- -1 per step, +10 food, +500 win, -500 die, 200 hit scared ghost
- Environment
- Actuators
- Sensors
Partially vs Fully Observable
- Fully: Agent can access the complete state of the world at each point in time (access to all info)
Deterministic vs Stochastic
- Deterministic: When we make a decision, does it cause certain actions to happen?
- Ex: Pacman moves up when we click up arrow
- Stochastic: There is some randomness, same set of outcomes not always guaranteed (env cannot be completely determined by the agent)
Static vs Non-Static
- Static: agent has time to compute decision
Discrete vs Continuous
- Continuous: continously operating controller
MEU

