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.
- Perceives environment
- through sensors
- acting upon it through actuators
Sensors produce percepts, creating a percept sequence.
#Types of Agents
- Simple reflex agents
- Model-based reflex agents
- Goal-based agents
- Learning agents
- Utility-based agents
#Sensors and Actuators
- Human
- Sensors: eyes, ears, etc.
- Actuators: hands, legs, etc.
- Software agent
- Sensors: file contents, typed input, etc.
- Actuators: writing files, playing sounds, etc.
#Agent Function
Agent function, mathematical map of inputs to outputs:
- Inputs
- Built-in knowledge
- Percept sequence
- Output
- Next action Agent programs (i.e. code or algorithm) is the way we would find it in practice.
#Performance Measure
Desirability of a sequence of environment states (caused by the sequence of actions of the agent)
“Do the right thing” = produce desired consequences (“consequentialism”)
- From the point of view of designer (not the machine)
- Can be explicit or sometimes implicit
- Should be about the outcome, not how to achieve it
- Can be fully described, or might learn from the user
Brief History