Download Multi-Agent System

INTELLIGENT SYSTEMS
Prof. Magdy M. Aboul-Ela
Information Systems Department
Faculty of Management and Information
Systems
French University in Egypt
Email: [email protected]
[email protected]
[email protected]
1
Textbook
S. Russell and P. Norvig Artificial
Intelligence: A Modern Approach Prentice
Hall, 2003, Second Edition
2
Outline
•
•
•
•
Course overview
What is AI?
A brief history
The state of the art
3
Course overview
•
•
•
•
•
•
Introduction and Agents
Search
Knowledge Representation
Symbolic Logic
Prolog
Applications
4
What is an
Artificial Intelligence (AI) ?
• AI attempts to build intelligent entities
• AI is both science and engineering:
– the science of understanding intelligent
entities — of developing theories which
attempt to explain and predict the nature of
such entities;
– the engineering of intelligent entities.
5
What is AI?
Views of AI fall into four categories:
Thinking humanly Thinking rationally
Acting humanly Acting rationally
6
Acting humanly: Turing Test
• Turing (1950) "Computing machinery and intelligence":
• "Can machines think?"  "Can machines behave intelligently?"
• Operational test for intelligent behavior: the Imitation Game
• Predicted that by 2000, a machine might have a 30% chance of
fooling a lay person for 5 minutes
• Anticipated all major arguments against AI in following 50 years
• Suggested major components of AI: knowledge, reasoning,
language understanding, learning
•
7
Thinking humanly: cognitive
modeling
• 1960s "cognitive revolution": informationprocessing psychology
•
• Requires scientific theories of internal activities
of the brain
•
• -- How to validate? Requires
1) Predicting and testing behavior of human subjects
(top-down)
or 2) Direct identification from neurological data
(bottom-up)
• Both approaches (roughly, Cognitive Science
and Cognitive Neuroscience) are now distinct
8
Thinking rationally: "laws of thought"
•
•
•
•
•
•
•
Aristotle: what are correct arguments/thought
processes?
Several Greek schools developed various forms of
logic: notation and rules of derivation for thoughts; may
or may not have proceeded to the idea of
mechanization
Direct line through mathematics and philosophy to
modern AI
Problems:
1.
2.
Not all intelligent behavior is mediated by logical deliberation
What is the purpose of thinking? What thoughts should I have?9
Acting rationally: rational agent
• Rational behavior: doing the right thing
•
• The right thing: that which is expected to
maximize goal achievement, given the
available information
•
• Doesn't necessarily involve thinking – e.g.,
blinking reflex – but thinking should be in
the service of rational action
10
Rational agents
• An agent is an entity that perceives and acts
• This course is about designing rational agents
•
• Abstractly, an agent is a function from percept
histories to actions:
•
[f: P*  A]
• For any given class of environments and tasks,
we seek the agent (or class of agents) with the
best performance
• Caveat: computational limitations make perfect
rationality unachievable
 design best program for given machine resources
11
AI prehistory
• Philosophy
• Mathematics
• Economics
• Neuroscience
• Psychology
• Computer
engineering
• Control theory
• Linguistics
Logic, methods of reasoning, mind as physical
system foundations of learning, language,
rationality
Formal representation and proof algorithms,
computation, (un)decidability, (in)tractability,
probability
utility, decision theory
physical substrate for mental activity
phenomena of perception and motor control,
experimental techniques
building fast computers
design systems that maximize an objective
function over time
knowledge representation, grammar
12
Abridged history of AI
•
•
•
•
•
1943
1950
1956
1952—69
1950s
• 1965
• 1966—73
•
•
•
•
•
1969—79
1980-1986-1987-1995--
McCulloch & Pitts: Boolean circuit model of brain
Turing's "Computing Machinery and Intelligence"
Dartmouth meeting: "Artificial Intelligence" adopted
Look, Ma, no hands!
Early AI programs, including Samuel's checkers
program, Newell & Simon's Logic Theorist,
Gelernter's Geometry Engine
Robinson's complete algorithm for logical reasoning
AI discovers computational complexity
Neural network research almost disappears
Early development of knowledge-based systems
AI becomes an industry
Neural networks return to popularity
AI becomes a science
The emergence of intelligent agents
13
State of the art
• Deep Blue defeated the reigning world chess champion
Garry Kasparov in 1997
• Proved a mathematical conjecture (Robbins conjecture)
unsolved for decades
• No hands across America (driving autonomously 98% of
the time from Pittsburgh to San Diego)
• During the 1991 Gulf War, US forces deployed an AI
logistics planning and scheduling program that involved
up to 50,000 vehicles, cargo, and people
• NASA's on-board autonomous planning program
controlled the scheduling of operations for a spacecraft
• Proverb solves crossword puzzles better than most
humans
14
Intelligent Agents
15
Outline
• Agents and environments
• Rationality
• PEAS (Performance measure,
Environment, Actuators, Sensors)
• Environment types
• Agent types
16
Agents
• An agent is anything that can be viewed as
perceiving its environment through sensors and
acting upon that environment through actuators
•
• Human agent: eyes, ears, and other organs for
sensors; hands,
• legs, mouth, and other body parts for actuators
•
• Robotic agent: cameras and infrared range
finders for sensors;
• various motors for actuators
17
•
Agents and environments
• The agent function maps from percept histories
to actions:
•
[f: P*  A]
• The agent program runs on the physical
architecture to produce f
18
Vacuum-cleaner world
• Percepts: location and contents, e.g.,
[A,Dirty]
•
• Actions: Left, Right, Suck, NoOp
•
19
Rational agents
• An agent should strive to "do the right thing",
based on what it can perceive and the actions it
can perform. The right action is the one that will
cause the agent to be most successful
•
• Performance measure: An objective criterion for
success of an agent's behavior
•
• E.g., performance measure of a vacuum-cleaner
agent could be amount of dirt cleaned up,
amount of time taken, amount of electricity
20
consumed, amount of noise generated, etc.
Rational agents
• Rational Agent: 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.
•
21
Rational agents
• Rationality is distinct from omniscience
(all-knowing with infinite knowledge)
•
• Agents can perform actions in order to
modify future percepts so as to obtain
useful information (information gathering,
exploration)
•
• An agent is autonomous if its behavior is
determined by its own experience (with
ability to learn and adapt)
22
PEAS
• PEAS: Performance measure, Environment,
Actuators, Sensors
• Must first specify the setting for intelligent agent
design
•
• Consider, e.g., the task of designing an
automated taxi driver:
• Performance measure
•
– Environment
– Actuators
– Sensors
23
PEAS
• Must first specify the setting for intelligent agent
design
•
• Consider, e.g., the task of designing an
automated taxi driver:
• Performance measure: Safe, fast, legal, comfortable trip,
maximize profits
•
– Environment: Roads, other traffic, pedestrians,
customers
– Actuators: Steering wheel, accelerator, brake, signal,
horn
– Sensors: Cameras, sonar, speedometer, GPS,
odometer, engine sensors, keyboard
24
PEAS
• Agent: Medical diagnosis system
• Performance measure: Healthy patient,
minimize costs, lawsuits
• Environment: Patient, hospital, staff
• Actuators: Screen display (questions,
tests, diagnoses, treatments, referrals)
•
• Sensors: Keyboard (entry of symptoms,
findings, patient's answers)
25
PEAS
• Agent: Part-picking robot
• Performance measure: Percentage of
parts in correct bins
• Environment: Conveyor belt with parts,
bins
• Actuators: Jointed arm and hand
• Sensors: Camera, joint angle sensors
26
PEAS
• Agent: Interactive English tutor
• Performance measure: Maximize student's
score on test
• Environment: Set of students
• Actuators: Screen display (exercises,
suggestions, corrections)
• Sensors: Keyboard
27
Environment types
• Fully observable (vs. partially observable): An agent's
sensors give it access to the complete state of the
environment at each point in time.
•
• Deterministic (vs. stochastic): The next state of the
environment is completely determined by the current
state and the action executed by the agent. (If the
environment is deterministic except for the actions of
other agents, then the environment is strategic)
•
• Episodic (vs. sequential): The agent's experience is
divided into atomic "episodes" (each episode consists of
the agent perceiving and then performing a single
action), and the choice of action in each episode
28
depends only on the episode itself.
Environment types
• Static (vs. dynamic): The environment is
unchanged while an agent is deliberating. (The
environment is semidynamic if the environment
itself does not change with the passage of time
but the agent's performance score does)
•
• Discrete (vs. continuous): A limited number of
distinct, clearly defined percepts and actions.
•
• Single agent (vs. multiagent): An agent
operating by itself in an environment.
29
Environment types
Fully observable
Deterministic
Episodic
Static
Discrete
Single agent
Chess with
a clock
Yes
Strategic
No
Semi
Yes
No
Chess without
a clock
Yes
Strategic
No
Yes
Yes
No
Taxi driving
No
No
No
No
No
No
• The environment type largely determines the agent design
•
• The real world is (of course) partially observable, stochastic,
sequential, dynamic, continuous, multi-agent
•
30
Agent functions and programs
• An agent is completely specified by the
agent function mapping percept
sequences to actions
• One agent function (or a small
equivalence class) is rational
•
• Aim: find a way to implement the rational
agent function concisely
•
31
Table-lookup agent
• Drawbacks:
– Huge table
– Take a long time to build the table
– No autonomy
– Even with learning, need a long time to learn
the table entries
32
Agent types
• Four basic types in order of increasing
generality:
•
• Simple reflex agents
• Model-based reflex agents
• Goal-based agents
• Utility-based agents
• Learning agents
33
Simple reflex agents
• Simple reflex agents act only on the basis
of the current percept. The agent function
is based on the condition-action rule: if
condition then action.
• This agent function only succeeds when
the environment is fully observable. Some
reflex agents can also contain information
on their current state which allows them to
disregard conditions whose actuators are
34
already triggered.
Simple reflex agents
35
Model-based agents
• Model-based agents can handle partially observable
environments. Its current state is stored inside the agent
maintaining some kind of structure which describes the
part of the world which cannot be seen.
• This behavior requires information on how the world
behaves and works. This additional information
completes the “World View” model.
• A model-based reflex agent keeps track of the current
state of the world using an internal model. It then
chooses an action in the same way as the reflex agent.
36
Model-based reflex agents
37
Goal-based agents
• Goal-based agents are model-based
agents which store information regarding
situations that are desirable.
• This allows the agent a way to choose
among multiple possibilities, selecting the
one which reaches a goal state.
38
Goal-based agents
39
Utility-based agents
• Goal-based agents only distinguish
between goal states and non-goal states.
It is possible to define a measure of how
desirable a particular state is.
• This measure can be obtained through the
use of a utility function which maps a state
to a measure of the utility of the state.
40
Utility-based agents
41
Learning agents
• Learning has an advantage that it allows
the agents to initially operate in unknown
environments and to become more
competent than its initial knowledge alone
might allow.
42
Learning agents
43
Other classes of intelligent agents
There are some of the sub-agents that may be a part of an Intelligent Agent or a
complete Intelligent Agent in themselves are:
•
•
•
•
•
•
•
•
Decision Agents (that are geared to decision making);
Input Agents (that process and make sense of sensor inputs - e.g. neural network
based agents);
Processing Agents (that solve a problem like speech recognition);
Spatial Agents (that relate to the physical real-world);
World Agents (that incorporate a combination of all the other classes of agents to
allow autonomous behaviors).
Believable agents - An agent exhibiting a personality via the use of an artificial
character (the agent is embedded) for the interaction.
Physical Agents - A physical agent is an entity which percepts through sensors and
acts through actuators.
Temporal Agents - A temporal agent may use time based stored information to
offer instructions or data acts to a computer program or human being and takes
program inputs percepts to adjust its next behaviors.
44
Multi-Agent System (MAS)
• A multi-agent system (MAS) is a system
composed of multiple interacting intelligent
agents.
• Multi-agent systems can be used to solve
problems which are difficult or impossible for an
individual agent or monolithic system to solve.
• Examples of problems which are appropriate to
multi-agent systems research include online
trading, disaster response, and modelling social
structures.
45
Multi-Agent System (MAS)
The agents in a multi-agent system have several important characteristics:
• Autonomy: the agents are at least partially autonomous
• Local views: no agent has a full global view of the system, or the system is
too complex for an agent to make practical use of such knowledge
• Decentralization: there is no designated controlling agent (or the system is
effectively reduced to a monolithic system)
•
•
•
Typically multi-agent systems research refers to software agents. However,
the agents in a multi-agent system could equally well be robots, humans or
human teams. A multi-agent system may contain combined human-agent
teams.
Multi-agent systems can manifest self-organization and complex behaviors
even when the individual strategies of all their agents are simple.
Agents can share knowledge using any agreed language, within the
constraints of the system's communication protocol. Example languages are
Knowledge Query Manipulation Language (KQML) or FIPA's Agent
46
Communication Language (ACL).
Solving problems by
searching
47
Outline
•
•
•
•
•
Problem-solving agents
Problem types
Problem formulation
Example problems
Basic search algorithms
48
Problem-solving agents
49
Example: Romania
• On holiday in Romania; currently in Arad.
• Flight leaves tomorrow from Bucharest
•
• Formulate goal:
– be in Bucharest
–
• Formulate problem:
– states: various cities
– actions: drive between cities
–
• Find solution:
– sequence of cities, e.g., Arad, Sibiu, Fagaras,
Bucharest
50
Example: Romania
51
Problem types
• Deterministic, fully observable  single-state problem
– Agent knows exactly which state it will be in; solution is a
sequence
–
• Non-observable  sensorless problem (conformant
problem)
– Agent may have no idea where it is; solution is a sequence
–
• Nondeterministic and/or partially observable 
contingency problem
– percepts provide new information about current state
– often interleave} search, execution
–
• Unknown state space  exploration problem
52
Example: vacuum world
• Single-state, start in #5.
Solution?
•
53
Example: vacuum world
• Single-state, start in #5.
Solution? [Right, Suck]
•
• Sensorless, start in
{1,2,3,4,5,6,7,8} e.g.,
Right goes to {2,4,6,8}
Solution?
•
54
Example: vacuum world
•
Sensorless, start in
{1,2,3,4,5,6,7,8} e.g.,
Right goes to {2,4,6,8}
Solution?
[Right,Suck,Left,Suck]
•
• Contingency
– Nondeterministic: Suck may
dirty a clean carpet
– Partially observable: location, dirt at current location.
– Percept: [L, Clean], i.e., start in #5 or #7
Solution?
55
Example: vacuum world
•
Sensorless, start in
{1,2,3,4,5,6,7,8} e.g.,
Right goes to {2,4,6,8}
Solution?
[Right,Suck,Left,Suck]
•
• Contingency
– Nondeterministic: Suck may
dirty a clean carpet
– Partially observable: location, dirt at current location.
– Percept: [L, Clean], i.e., start in #5 or #7
Solution? [Right, if dirt then Suck]
56
Single-state problem formulation
A problem is defined by four items:
1. initial state e.g., "at Arad"
2. actions or successor function S(x) = set of action–state pairs
– e.g., S(Arad) = {<Arad  Zerind, Zerind>, … }
–
3. goal test, can be
– explicit, e.g., x = "at Bucharest"
– implicit, e.g., Checkmate(x)
–
4. path cost (additive)
– e.g., sum of distances, number of actions executed, etc.
– c(x,a,y) is the step cost, assumed to be ≥ 0
–
•
A solution is a sequence of actions leading from the initial state to a
57
goal state
Selecting a state space
• Real world is absurdly complex
 state space must be abstracted for problem solving
• (Abstract) state = set of real states
• (Abstract) action = complex combination of real actions
– e.g., "Arad  Zerind" represents a complex set of possible
routes, detours, rest stops, etc.
• For guaranteed realizability, any real state "in Arad“ must
get to some real state "in Zerind"
• (Abstract) solution =
– set of real paths that are solutions in the real world
• Each abstract action should be "easier" than the original
problem
58
Vacuum world state space graph
•
•
•
•
•
states?
actions?
goal test?
path cost?
59
Vacuum world state space graph
•
•
•
•
states? integer dirt and robot location
actions? Left, Right, Suck
goal test? no dirt at all locations
path cost? 1 per action
60
Example: The 8-puzzle
•
•
•
•
states?
actions?
goal test?
path cost?
61
Example: The 8-puzzle
•
•
•
•
•
states? locations of tiles
actions? move blank left, right, up, down
goal test? = goal state (given)
path cost? 1 per move
[Note: optimal solution of n-Puzzle family is NP-hard]
62
Example: robotic assembly
• states?: real-valued coordinates of robot joint
angles parts of the object to be assembled
•
• actions?: continuous motions of robot joints
• goal test?: complete assembly
• path cost?: time to execute
63
Tree search algorithms
• Basic idea:
– offline, simulated exploration of state space by
generating successors of already-explored states
(a.k.a.~expanding states)
–
64
Tree search example
65
Tree search example
66
Tree search example
67
Implementation: general tree search
68
Implementation: states vs. nodes
• A state is a (representation of) a physical configuration
• A node is a data structure constituting part of a search
tree includes state, parent node, action, path cost g(x),
depth
• The Expand function creates new nodes, filling in the
various fields and using the SuccessorFn of the
problem to create the corresponding states.
•
69
Search strategies
• A search strategy is defined by picking the order of node
expansion
• Strategies are evaluated along the following dimensions:
–
–
–
–
–
completeness: does it always find a solution if one exists?
time complexity: number of nodes generated
space complexity: maximum number of nodes in memory
optimality: does it always find a least-cost solution?
• Time and space complexity are measured in terms of
– b: maximum branching factor of the search tree
– d: depth of the least-cost solution
– m: maximum depth of the state space (may be ∞)
–
70
Uninformed search strategies
• Uninformed search strategies use only the
information available in the problem
definition:
– Breadth-first search
– Uniform-cost search
– Depth-first search
– Depth-limited search
– Iterative deepening search
71
Breadth-first search
• Expand shallowest unexpanded node
• Implementation:
– fringe is a FIFO queue, i.e., new successors
go at end
–
72
Breadth-first search
• Expand shallowest unexpanded node
• Implementation:
– fringe is a FIFO queue, i.e., new successors
go at end
–
73
Breadth-first search
• Expand shallowest unexpanded node
•
• Implementation:
– fringe is a FIFO queue, i.e., new successors
go at end
–
74
Breadth-first search
• Expand shallowest unexpanded node
• Implementation:
– fringe is a FIFO queue, i.e., new successors
go at end
–
75
Properties of breadth-first search
•
•
•
•
•
•
•
•
Complete? Yes (if b is finite)
Time? 1+b+b2+b3+… +bd + b(bd-1) = O(bd+1)
Space? O(bd+1) (keeps every node in memory)
Optimal? Yes (if cost = 1 per step)
• Space is the bigger problem (more than time)
76
Uniform-cost search
• Expand least-cost unexpanded node
•
• Implementation:
– fringe = queue ordered by path cost
• Equivalent to breadth-first if step costs all equal
•
• Complete? Yes, if step cost ≥ ε
• Time? # of nodes with g ≤ cost of optimal solution,
O(bceiling(C*/ ε)) where C* is the cost of the optimal solution
• Space? # of nodes with g ≤ cost of optimal solution,
O(bceiling(C*/ ε))
• Optimal? Yes – nodes expanded in increasing order of
g(n)
77
•
Depth-first search
• Expand deepest unexpanded node
• Implementation:
– fringe = LIFO queue, i.e., put successors at front
–
78
Depth-first search
• Expand deepest unexpanded node
•
• Implementation:
– fringe = LIFO queue, i.e., put successors at front
–
79
Depth-first search
• Expand deepest unexpanded node
• Implementation:
– fringe = LIFO queue, i.e., put successors at front
–
80
Depth-first search
• Expand deepest unexpanded node
• Implementation:
– fringe = LIFO queue, i.e., put successors at front
–
81
Depth-first search
• Expand deepest unexpanded node
• Implementation:
– fringe = LIFO queue, i.e., put successors at front
–
82
Depth-first search
• Expand deepest unexpanded node
• Implementation:
– fringe = LIFO queue, i.e., put successors at front
–
83
Depth-first search
• Expand deepest unexpanded node
• Implementation:
– fringe = LIFO queue, i.e., put successors at front
–
84
Depth-first search
• Expand deepest unexpanded node
• Implementation:
– fringe = LIFO queue, i.e., put successors at front
–
85
Depth-first search
• Expand deepest unexpanded node
• Implementation:
– fringe = LIFO queue, i.e., put successors at front
–
86
Depth-first search
• Expand deepest unexpanded node
• Implementation:
– fringe = LIFO queue, i.e., put successors at front
–
87
Depth-first search
• Expand deepest unexpanded node
• Implementation:
– fringe = LIFO queue, i.e., put successors at front
–
88
Depth-first search
• Expand deepest unexpanded node
• Implementation:
– fringe = LIFO queue, i.e., put successors at front
–
89
Properties of depth-first search
• Complete? No: fails in infinite-depth spaces,
spaces with loops
– Modify to avoid repeated states along path
–
 complete in finite spaces
• Time? O(bm): terrible if m is much larger than d
– but if solutions are dense, may be much faster than
breadth-first
• Space? O(bm), i.e., linear space!
• Optimal? No
90
Depth-limited search
= depth-first search with depth limit l,
i.e., nodes at depth l have no successors
• Recursive implementation:
91
Iterative deepening search
92
Iterative deepening search l =0
93
Iterative deepening search l =1
94
Iterative deepening search l =2
95
Iterative deepening search l =3
96
Iterative deepening search
• Number of nodes generated in a depth-limited search to
depth d with branching factor b:
NDLS = b0 + b1 + b2 + … + bd-2 + bd-1 + bd
• Number of nodes generated in an iterative deepening
search to depth d with branching factor b:
NIDS = (d+1)b0 + d b^1 + (d-1)b^2 + … + 3bd-2 +2bd-1 + 1bd
• For b = 10, d = 5,
– NDLS = 1 + 10 + 100 + 1,000 + 10,000 + 100,000 = 111,111
–
– NIDS = 6 + 50 + 400 + 3,000 + 20,000 + 100,000 = 123,456
–
• Overhead = (123,456 - 111,111)/111,111 = 11%
97
Properties of iterative
deepening search
• Complete? Yes
•
• Time? (d+1)b0 + d b1 + (d-1)b2 + … + bd =
O(bd)
•
• Space? O(bd)
•
• Optimal? Yes, if step cost = 1
98
Summary of algorithms
99
Repeated states
• Failure to detect repeated states can turn
a linear problem into an exponential one!
•
100
Summary
• Problem formulation usually requires abstracting away
real-world details to define a state space that can
feasibly be explored
•
• Variety of uninformed search strategies
•
• Iterative deepening search uses only linear space and
not much more time than other uninformed algorithms
•
101