Chapter 2.1 - 2.3
Agents and Environments
- Agent
- Environment
- Sensors (percept, percept sequence)
- Actuators (actions)
- Figure 2.1
- An agent function maps a percept sequence to an action
- Vacuum World: Figures 2.2, 2.3
Good Behavior: The Concept of Rationality
- Performance measure: it is better to design the performance measure
according to what one actually wants in the environment, rather
than according to how one thinks the agent should behave
- Rationality depends on performance measure, prior knowledge,
performable actions, percept sequence
- Rational agent: for each possible percept sequence, a rational
agent should select an action that is expected to maximize its
performance measure, given the evidence provided by the percept
sequence and whatever built-in knowledge the agent has
- Rationality does not mean omniscience - an agent tries to maximize its expected
performance, not its actual performance
- Agents can engage in information gathering, exploration and
learning
Task Environments
- PEAS: performance measure, environment, actuators, sensors
- Figures 2.4 and 2.5
Task Environment Properties
- Fully observable vs. partially observable
- Deterministic vs. stochastic
- Episodic vs. sequential
- Static vs. dynamic
- Discrete vs. continuous
- Single agent vs. multiagent
- Figure 2.6
Agent Examples