intelligent agents

20
Intelligent Agents

Upload: yehuda

Post on 25-Feb-2016

54 views

Category:

Documents


0 download

DESCRIPTION

Intelligent Agents. Software agents. Monday: Overview video (Introduction to software agents) Agents and environments Rationality Wednesday: PEAS (Performance measure, Environment, Actuators, Sensors ) Example of simple and not-so-simple software agents Friday: Environment types - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Intelligent Agents

Intelligent Agents

Page 2: Intelligent Agents

Software agentsO Monday:

O Overview video (Introduction to software agents)O Agents and environmentsO Rationality

O Wednesday:O PEAS (Performance measure, Environment, Actuators,

Sensors)O Example of simple and not-so-simple software agents

O Friday:O Environment typesO Agent typesO Group discussion

Page 3: Intelligent Agents

Very short software agents overview

O http://labcast.media.mit.edu/?p=23

O What is an agent?: A computer program that is O Embodied (i.e local)O Situated (ie. has knowledge of / responds to

its situation / environment)O Autonomous (i.e decides what action to take

based on its own situation)O Cooperating (i.e can communicate with other

agents to achieve tasks)

Page 4: Intelligent Agents

Agent and environments

Page 5: Intelligent Agents

Agent and environments

O Agents include humans, robots, softbots, thermostats, etc.

O The agent function maps from percept histories to actions:

f : P -> AO The agent program runs on the

physical architecture to produce f

Page 6: Intelligent Agents

Vacuum-cleaner world

O Percepts: location and contents, e.g., [A;Dirty]

O Actions: Left, Right, Suck, NoOp

Page 7: Intelligent Agents

A vacuum-cleaner agent

Page 8: Intelligent Agents

Pseudocode

What is the right function?Can it be implemented in a small agent program?

Page 9: Intelligent Agents

RationalityO Fixed performance measure evaluates the environment

sequenceO one point per square cleaned up in time T?O one point per clean square per time step, minus one per move?O penalize for > k dirty squares?

O A rational agent chooses whichever action maximizes the expected value of the performance measure given the percept sequence to date

O Rational is different from omniscientO percepts may not supply all relevant information

O Rational is different from clairvoyantO action outcomes may not be as expected

O Rational is different from successfulO Rational exploration, learning, autonomy

Page 10: Intelligent Agents

PEASO To design a rational agent, we must

specify the task environment.Example: consider, e.g., the task of designing an automated taxi:O Performance measure: O Environment:O Actuators: O Sensors:

Page 11: Intelligent Agents

Another exampleO Internet shopping agents:

O Performance measure??O Environment??O Actuators??O Sensors??

Page 12: Intelligent Agents

Environment types

O The environment type largely determines the agent design

O The real world is (of course) partially observable, stochastic, sequential, dynamic, continuous, multi-agent

Page 13: Intelligent Agents

Agent typesO Four basic types in order of

increasing generality:O Simple reflex agentsO Reflex agents with stateO Goal-based agentsO Utility-based agents

Page 14: Intelligent Agents

Simple reflex agents

Page 15: Intelligent Agents

Reflex agents with state

Page 16: Intelligent Agents

Goal-based agents

Page 17: Intelligent Agents

Utility-based agents

Page 18: Intelligent Agents

Learning agents

Page 19: Intelligent Agents

ExampleO Simple example

O Demonstration of software agent on Linux

O Complex example:O http

://web.media.mit.edu/~hugo/demos/commonsense-aria.html

Page 20: Intelligent Agents

SummaryO Agents interact with environments through actuators and

sensorsO The agent function describes what the agent does in all

circumstancesO The performance measure evaluates the environment

sequenceO A perfectly rational agent maximizes expected performanceO Agent programs implement (some) agent functionsO PEAS descriptions define task environmentsO Environments are categorized along several dimensions:O observable? deterministic? episodic? static? discrete? single-

agent?O Several basic agent architectures exist:

O Reflex, reflex with state, goal-based, utility-based