Scribd
Upload
20 views

m2 Agents

This document provides an outline of key concepts about intelligent agents, including: 1) It defines agents and environments, describing how agents perceive their environment through sensors and act through actuators. 2) It introduces the PEAS framework for defining agents, which describes the Performance measure, Environment, Actuators, and Sensors for an agent. 3) It describes different types of environments an agent can operate in, including fully observable vs. partially observable, deterministic vs. stochastic, and more.

Uploaded by

Dr. MNV Kiranbabu
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
20 views

m2 Agents

This document provides an outline of key concepts about intelligent agents, including: 1) It defines agents and environments, describing how agents perceive their environment through sensors and act through actuators. 2) It introduces the PEAS framework for defining agents, which describes the Performance measure, Environment, Actuators, and Sensors for an agent. 3) It describes different types of environments an agent can operate in, including fully observable vs. partially observable, deterministic vs. stochastic, and more.

Uploaded by

Dr. MNV Kiranbabu
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 28

Intelligent Agents

Chapter 2
Outline

Agents and environments

Rationality

PEAS (Performance measure


Environment Actuators Sensors!

Environment types

Agent types
Agents

An agent is anything that can "e vie#ed as


perceiving its environment through sensors and
acting upon that environment through actuators

$uman agent% eyes ears and other organs for


sensors& hands

legs mouth and other "ody parts for actuators

Ro"otic agent% cameras and infrared range


finders for sensors&

various motors for actuators

Agents and environments

'he agent function maps from percept histories


to actions%
(f% P* A)

'he agent program runs on the physical


architecture to produce f

agent * architecture + program

,acuum-cleaner #orld

Percepts% location and contents e.g.


(A/irty)

Actions% Left Right Suck NoOp

A vacuum-cleaner agent

0input1ta"les2vacuum-agent-function-ta"le3

Rational agents

An agent should strive to 4do the right thing4


"ased on #hat it can perceive and the actions it
can perform. 'he right action is the one that #ill
cause the agent to "e most successful

Performance measure% An o"5ective criterion for


success of an agent6s "ehavior

E.g. performance measure of a vacuum-cleaner


agent could "e amount of dirt cleaned up
amount of time ta7en amount of electricity
consumed amount of noise generated etc.

Rational agents

Rational Agent% 8or each possi"le percept


se9uence a rational agent should select
an action that is e:pected to ma:imi;e its
performance measure given the evidence
provided "y the percept se9uence and
#hatever "uilt-in 7no#ledge the agent
has.

Rational agents

Rationality is distinct from omniscience


(all-7no#ing #ith infinite 7no#ledge!

Agents can perform actions in order to


modify future percepts so as to o"tain
useful information (information gathering
e:ploration!

An agent is autonomous if its "ehavior is


determined "y its o#n e:perience (#ith
a"ility to learn and adapt!

PEAS

PEAS% Performance measure Environment


Actuators Sensors

<ust first specify the setting for intelligent agent


design

Consider e.g. the tas7 of designing an


automated ta:i driver%
=
Performance measure
=
Environment
=
Actuators
=
Sensors
=
=

PEAS

<ust first specify the setting for intelligent agent


design

Consider e.g. the tas7 of designing an


automated ta:i driver%
=
Performance measure% Safe fast legal comforta"le
trip ma:imi;e profits
=
Environment% Roads other traffic pedestrians
customers
=
Actuators% Steering #heel accelerator "ra7e signal
horn
=
Sensors% Cameras sonar speedometer >PS
odometer engine sensors 7ey"oard
=
=
=
=

PEAS

Agent% <edical diagnosis system

Performance measure% $ealthy patient


minimi;e costs la#suits

Environment% Patient hospital staff

Actuators% Screen display (9uestions


tests diagnoses treatments referrals!

Sensors% ?ey"oard (entry of symptoms


findings patient6s ans#ers!

PEAS

Agent% Part-pic7ing ro"ot

Performance measure% Percentage of


parts in correct "ins

Environment% Conveyor "elt #ith parts


"ins

Actuators% @ointed arm and hand

Sensors% Camera 5oint angle sensors


PEAS

Agent% Interactive English tutor

Performance measure% <a:imi;e


student6s score on test

Environment% Set of students

Actuators% Screen display (e:ercises


suggestions corrections!

Sensors% ?ey"oard
Environment types

8ully o"serva"le (vs. partially o"serva"le!% An agent6s


sensors give it access to the complete state of the
environment at each point in time.

/eterministic (vs. stochastic!% 'he ne:t state of the


environment is completely determined "y the current
state and the action e:ecuted "y the agent. (If the
environment is deterministic e:cept for the actions of
other agents then the environment is strategic!

Episodic (vs. se9uential!% 'he agent6s e:perience is


divided into atomic 4episodes4 (each episode consists of
the agent perceiving and then performing a single
action! and the choice of action in each episode
depends only on the episode itself.

Environment types

Static (vs. dynamic!% 'he environment is


unchanged #hile an agent is deli"erating. ('he
environment is semidynamic if the environment
itself does not change #ith the passage of time
"ut the agent6s performance score does!

/iscrete (vs. continuous!% A limited num"er of


distinct clearly defined percepts and actions.

Single agent (vs. multiagent!% An agent operating


"y itself in an environment.

Environment types
Chess #ith Chess #ithout 'a:i driving
a cloc7 a cloc7
8ully o"serva"le Aes Aes Bo
/eterministic Strategic Strategic Bo
Episodic Bo Bo Bo
Static Semi Aes Bo
/iscrete Aes Aes Bo
Single agent Bo Bo Bo
'he environment type largely determines the agent design
'he real #orld is (of course! partially o"serva"le stochastic
se9uential dynamic continuous multi-agent

Agent functions and programs

An agent is completely specified "y the


agent function mapping percept
se9uences to actions

One agent function (or a small


e9uivalence class! is rational

Aim% find a #ay to implement the rational


agent function concisely

'a"le-loo7up agent

0input1algorithms2ta"le-agent-algorithm3

/ra#"ac7s%
=
$uge ta"le
=
'a7e a long time to "uild the ta"le
=
Bo autonomy
=
Even #ith learning need a long time to learn
the ta"le entries

Agent program for a vacuum-


cleaner agent

0input1algorithms2refle:-vacuum-agent-
algorithm3

Agent types

8our "asic types in order of increasing


generality%

Simple refle: agents

<odel-"ased refle: agents

>oal-"ased agents

Ctility-"ased agents

Simple refle: agents


Simple refle: agents

0input1algorithms2d-agent-algorithm3

<odel-"ased refle: agents


<odel-"ased refle: agents

0input1algorithms2d+-agent-algorithm3

>oal-"ased agents

Ctility-"ased agents
Dearning agents

You might also like