MULTI AGENT SYSTEMS

 Distributed Artificial Intelligence DAI Subfield of AI  Development of distributed solutions for complex problems  problem that is beyond the capability of an individual problem s

 [1] Michael Wooldridge, “An Introduction to MultiAgent Systems”,

Second Edition, 2009

 [2] R.H Bordini, J.F.Hubner, M Wooldridge, “Programming

multi-agent systems in AgentSpeak using Jason”, 2007.

 Background

 Agent

 Environment

 Architecture for Agents

Reading: Chapter 1&2, [1]

 Distributed Artificial Intelligence (DAI)

 Subfield of AI

 Development of distributed solutions for complex problems

 problem that is beyond the capability of an individual problem

solver

 Two mainstreams

 Distributed prolem solving (DPS)

 MultiAgent systems (MAS)

 Distributed Artificial Intelligence (DAI)

 Two mainstreams

 Distributed prolem solving (DPS)

 Centralized Control, Distributed Data

 MultiAgent systems (MAS)

 Distributed Control, Distributed Data

 a system comprising several agents that “live” and interact in the same

environment

 Cleaning robot

 Gold miners

 There is no universally accepted definition of the term “Agent”

 There is a general consensus that autonomy is central to the

notion of agency

 Difficulty is that various attributes associated with agency are

of diffening importance for different domains

 Autonomy:

 capable of acting independently,

 exhibiting control over their internal state

 Thus: an agent is a computer system capable of autonomous action

in some environment in order to meet its design objectives

SYSTEM

ENVIRONMENT

 An agent in its environment

Sensors

Feedback

Actions

 In most domain of reasonable complexity, an agent will not have

complete control over its environment

  It will have at best partial control, in that it can influence it

 Trivial (non-interesting) agents:

 Thermostat

 Have a sensor for detecting room temperature

 Two signals: too low, and OK

 Available actions: heating on , and heating off

 Rules:

 Too cold  heating on

 Temperature Ok  heating off

 When the door of the room is close?  guaranteed effects

 When the door of the room is open?

 An intelligent agent is a computer system capable of flexible

autonomous action in some environment

 By flexible, we mean:

 reactive

 pro-active

 social

 If a program’s environment is guaranteed to be fixed, the program

need never worry about its own success or failure – program just executes blindly

 Example of fixed environment: compiler

incomplete Many (most?) interesting environments are dynamic

 A reactive system is one that maintains an ongoing interaction

with its environment, and responds to changes that occur in it (in time for the response to be useful)

 we generally want agents to do things for us

 goal directed behavior

 Pro-activeness = generating and attempting to achieve goals;

not driven solely by events; taking the initiative

 Recognizing opportunities

 The real world is a multi-agent environment: we cannot go around attempting

to achieve goals without taking others into account

 Some goals can only be achieved with the cooperation of others

 Social ability in agents is the ability to interact with other agents (and possibly

humans) via some kind of agent-communication language, and perhaps

cooperate with others

 An accessible environment is one in which the agent can obtain complete,

accurate, up-to-date information about the environment’s state

 Most moderately complex environments (including, for example, the

everyday physical world and the Internet) are inaccessible

Accessible vs inaccessible

 The more accessible an environment is, the simpler it is to build

agents to operate in it

 Example:

 a vacuum agent with only a local dirt sensor cannot tell whether there is

dirt in other squares,

 an automated taxi cannot see what other drivers are thinking

 A deterministic environment is one in which any action has a single

guaranteed effect — there is no uncertainty about the state that will result from performing an action

Deterministic vs non-deterministic

 Example:

 The vacuum world as we described it is deterministic, but variations

can include stochastic elements such as randomly appearing dirt and an unreliable suction mechanism

 Taxi driving is clearly non-deterministic in this sense, because one

can never predict the behavior of traffic exactly; moreover, one's tires blow out and one's engine seizes up without warning

 A static environment is one that can be assumed to remain

unchanged except by the performance of actions by the agent

 A dynamic environment is one that has other processes

operating on it, and which hence changes in ways beyond the agent’s control

Static vs dynamic

 Other processes can interfere with the agent’s actions (as in

concurrent systems theory)

 The physical world is a highly dynamic environment

 Example:

 Taxi driving is clearly dynamic

 Crossword puzzles are static

 An environment is discrete if there are a fixed, finite number

of actions and percepts in it

 Russell and Norvig give a chess game as an example of a

discrete environment, and taxi driving as an example of a continuous one

Discrete vs continuous

 Example:

 a discrete-state environment such as a chess game has a finite

number of distinct states Chess also has a discrete set of percepts and actions

 Taxi driving is a continuous- state and continuous-time problem:

the speed and location of the taxi and of the other vehicles sweep through a range of continuous values and do so smoothly over time

 BDI = Belief – Desire - Intention

belief revision

generate options

act sense

Environment

 Desires and intentions are realized from plan library

 Cleaning agent