, AGENTS & ENVIRONMENTS
Agent Environment
Sens Ors
, <
percepts
Acu -Éators Actions
>
An
agent PERCIEVES its environment through SENSORS { acts upon
•
it throughACUTAT0R Actuators * *
•
Ex .
→ Humans are
agents
vision , audio touch smell taste
•
sensors =
, ,
,
Acutators brain state
muscles, secretions ,
changing
• =
→ Pocket calculators are
agents
•
sensors =
key state sensors
•
Acutators =
digit display
•
Al is more interested in agents with large computational resources
& environments that require non trivial decision making -
•
Agent 's percept sequence = complete history of the
everything
agent has ever percieued
•
Agent 's depend on its built in knowledge
choice of actions -
{ on the entire percept sequence observed to date [not on
anything
it hasn't percieved ]
AGENT FUNCTIONS
•
The AGENT FUNCTION maps from percept histories to actions
•
f :p *
→ A
→
Agent 's actual response to any sequence of percepts
AGENT PROGRAMS
•
The AGENT PROGRAM 1 runs on some machine M to implement f
f Agent (I , M)
• =
Real machines have limited speed { memory , delay
→
introducing ,
so
agent function f depends on M as well as I
Agent Environment
Sens Ors
, <
percepts
Acu -Éators Actions
>
An
agent PERCIEVES its environment through SENSORS { acts upon
•
it throughACUTAT0R Actuators * *
•
Ex .
→ Humans are
agents
vision , audio touch smell taste
•
sensors =
, ,
,
Acutators brain state
muscles, secretions ,
changing
• =
→ Pocket calculators are
agents
•
sensors =
key state sensors
•
Acutators =
digit display
•
Al is more interested in agents with large computational resources
& environments that require non trivial decision making -
•
Agent 's percept sequence = complete history of the
everything
agent has ever percieued
•
Agent 's depend on its built in knowledge
choice of actions -
{ on the entire percept sequence observed to date [not on
anything
it hasn't percieved ]
AGENT FUNCTIONS
•
The AGENT FUNCTION maps from percept histories to actions
•
f :p *
→ A
→
Agent 's actual response to any sequence of percepts
AGENT PROGRAMS
•
The AGENT PROGRAM 1 runs on some machine M to implement f
f Agent (I , M)
• =
Real machines have limited speed { memory , delay
→
introducing ,
so
agent function f depends on M as well as I
