Agent and the Knowledge Representation-6_411

download Agent and the Knowledge Representation-6_411

of 10

Transcript of Agent and the Knowledge Representation-6_411

  • 8/7/2019 Agent and the Knowledge Representation-6_411

    1/10

  • 8/7/2019 Agent and the Knowledge Representation-6_411

    2/10

    ability to reuse methods in different branches of

    the hierarchy.

    The authority for decision making and

    control in concentrated in a single problem solver

    at each level of hierarchy . Interaction is through

    vertical communication from superior to sub-

    ordinate agent and vice-versa . [3,4]

    2.2 Modular Organization

    A modular organization is one in which

    different functional components are separated

    from one another, a technique adopted from

    software engineering. This is in contrast to a

    composite organization in which there is no

    separation between functions. Modular

    organization is also distinct from hierarchical

    organization. Modular organization is chiefly

    concerned with the horizontal design of a system

    whereas hierarchical organization involves a

    consideration of the vertical nature of the design.

    Thus, each level in a hierarchical system may be

    sub-divided into functionally-distinct modules.

    3 KNOWLEDGE REPRESENTATION

    Knowledge means intellectual acquaintance with

    perception of fact or truth. Representation is a

    way of describing certain fragments or

    information so that any reasoning system can

    easily adopt it for inferencing purposes.

    Knowledge Representation system should

    provide way of representing complex knowledge

    and should possess the following characteristics:

    1. The representation scheme should have a set

    of well defined syntax and semantics. Thiswill help in representing various kinds of

    knowledge .

    2. The knowledge representation scheme shouldhave a good expressive capacity . A good

    expressive capacity will catalyze the

    inferencing mechanism in its reasoning

    process.The philosophy behind Subsumption Architecture

    is that the world should be its own model.

    According to Brooks, storing models of the world

    is dangerous in dynamic, unpredictable

    environments because representations might be

    incorrect or outdated. What is needed is the abilityto react quickly to the present. Thus, sensor

    readings should be mapped quickly and directly to

    actuator commands in a decentralized fashion.

    Agents designed according to Subsumption

    Architecture are generally non-symbolic. They

    have no global representation, and are

    decentralized.

    In ATALANTIS architecture the control

    layer uses procedural knowledge while the

    deliberative layer uses declarative knowledge.

    THEO is frame based architecture .For an

    architecture to be frame-based means that it is

    organized around the representation of all of its

    knowledge in frames. This is important because it

    allows a uniform representation of knowledge. In

    a frame based system, all knowledge is

    represented as slots of some concept which

    describe properties of that concept

    All knowledge in PRODIGY including

    control rules, domain knowledge, learned

    knowledge is represented in the PRODIGY

    description language (PDL). PDL is based on

    first-order predicate calculus and includes

    variables, disjunction and conjunction, universal

    and existential quantification, and conditional

    representation.

    In Icarus, the Concept tree in Labyrinth stores

    all knowledge. The other components of the

    architecture have access to this knowledge

    explicitly. For example, Daedalus retrieves

    explicit plans from Labyrinth. The use ofdeclarative representations allows each module to

    reason about the available knowledge . [8]

    The symbolic world model or global database

    in AIS is formed from a uniform and declarative

    based conceptual network. Within this

    representation, sub-networks can represent

    actions, events, and control plans through

    symbolic abstraction. The generality of the

    conceptual network representation permits

    encoding complete knowledge from different

    problem domains, enhancing multi-faceted

    expertise. The network-type representation

    facilitates explanation as well .All knowledge in MAX agents exists in

    nested logic frames including the rules and

    operators. Separating rules and operators

    integrates a domain theory, in which rules specify

    actions and actions satisfy rules. With a

    production memory architecture knowledge is

    brought to bear on a problem by executing some

    operator that uses that knowledge, i.e. it adds

    symbols and it deletes symbols in some logical or

    rational fashion

    Homer uses both Black-box representation

    and heterogeneous representation .SOAR uses

    Black-box representation .Teton , RALPH-MEAand Entropy Reduction Engine architectures use

    both Declarative and Procedural representation .

    3.1 Symbolic World Model

    For some of the architectures under consideration ,

    there (usually) exists a level that enables the

    system to store knowledge in some basic

    framework. This framework is based on symbols

    which serve to represent relations between the

    agent and its environment, and hence knowledge,

    within the system. Such symbolic abstraction

  • 8/7/2019 Agent and the Knowledge Representation-6_411

    3/10

    occurs at the symbol level and derives its power

    from the physical symbol system hypothesis.

    Thus, symbolic abstraction enables, or at least

    facilitate, many architecture capabilities including

    planning and learning. Soar's productions use

    symbols to store information . Such systems have

    memory ,symbols ,operations ,interpretation and

    capacities to operate on symbols. [4]

    3.2 Size Of The Knowledge Base

    There are physical limits to the size of the

    knowledge base that may be supported by an

    individual agent. But, in addition to the physical

    limit on memory , there may be limitations

    imposed by the agent's architecture and style of

    control. For example, Homer experiences a

    processing slow-down as its episodic knowledge

    base increases in size. Since productions can be

    easily added to the knowledge base or learned in

    Soar, the knowledge base may grow incrementally

    to cover a wider range of behaviors. The ultimatesize of the knowledge base is dictated only by the

    desired functionality; no absolute limits are made

    by the architecture or the representation.

    3.3 Glass Box Approach

    Glass box knowledge representation may be

    defined as the ability of rules to examine each

    other. Glass box representation is useful for

    modular architectures, so that all modules have

    access to all knowledge. Glass box representation

    allows the rules and the architecture to share

    responsibility to examine, activate and rewards

    other rules. Also, learning is facilitated by glass

    box representation because learning modules may

    write. Related properties are uniform access to

    knowledge and homogeneous representation .

    Knowledge in Prodigy is stored in a global

    database which is freely available to all other

    modules. Knowledge is represented uniformly to

    simplify access by other modules. The declarative

    representation of knowledge allows the

    architecture to analyze its own actions and

    decisions.

    3.4 Black Box Approach

    Black box knowledge representation may be

    defined as the inability of rules to examine other

    rules. Architectures with this property are limited

    in their direct inferencing ability to see what other

    rules are up to. Such representations are less

    fragile to changes and more modular and easier to

    understand.

    Black-box knowledge representation does not rule

    out meta-reasoning, but it would make it more

    circuituitous. Some knowledge is shared amongst

    all execution architectures, such as possible states

    and actions. However, each architecture also

    requires its own specific knowledge. This

    knowledge includes utility values of actions and

    states, which are not shared amongst architectures

    in part because each architecture needs a different

    type of utility value. Each execution architecture

    operates independently and in parallel with the

    others, with no interaction of the knowledge of

    different architectures occurring. Each individual

    execution architecture does not have its own meta-

    knowledge in regards to what it knows. RALPH-

    MEA uses this technique .

    3.5 Declarative Representation

    A declarative representation declares every

    piece of knowledge and permits the reasoning

    system to use rules of inference to come out with

    a new piece of information .A classic example of a

    declarative representation is logic. The primaryadvantages of declarative knowledge is the ability

    to use knowledge in ways that the system designer

    did not foresee. Often times, whether knowledge

    is viewed as declarative or procedural is not an

    intrinsic property of the knowledge base, but is a

    function of what is allowed to read from it.

    Production systems, for example, are declarative if

    productions may view themselves, and are

    procedural it they cannot. (cf. glass-box/black-box

    control knowledge).

    A particular architecture may use both

    declarative and procedural knowledge at different

    times, taking advantage of their differentadvantages. The architectures that use Declarative

    representation are ERE , ICARUS , MAX ,

    Prodigy and THEO . ERE uses Uniform

    conceptual graph to store information . Homer ,

    ICARUS , MAX and Theo use Frames . The

    Prodigy use Prodigy description language .

    3.7 Procedural Representation

    Architectures with procedural representations

    encode how to do some task. In other words,

    procedural knowledge is skill knowledge. One

    advantage of procedural representations is

    possibly faster usage in a performance system.

    Productions are a common means of representingprocedural knowledge.

    Use of procedural knowledge in an agent

    raises the questions of whether the agent can

    "know what it knows" . That the agent can

    demonstrate that it "knows what it knows" is

    illustrated in a Soar system which includes the

    ability to explain its actions. Procedural

    representation is where knowledge is intrinsically

    bound up in the routines and procedures which use

    it . These procedures and routines know how to do

  • 8/7/2019 Agent and the Knowledge Representation-6_411

    4/10

    a particular task which would be regarded as

    intelligent .

    3.8 Global Representation and Uniform Access to

    Knowledge

    In architectures with global knowledge,

    different modules may read and/or write to

    common database(s). Often this knowledge is

    used to representing a world-view of what the

    agent believes is true in its environment. An

    advantage of having global knowledge is that

    different modules may share their data and

    abilities for more intelligent combined behavior.

    This makes modular architectures more effective.

    Also, such knowledge is necessary for

    representing world-views: without it an

    architecture may only react to its present sensor

    readings (e.g. Subsumption Architecture.)[9].

    A related property is knowledge

    homogeneity, which is a measure of how similar

    knowledge is represented by the architecture.

    Homer's memory modules contains stateinformation which is globally available. The

    planning module retrieves goals from the generic

    memory. The interpreter and perception modules

    enter events into memory. Although the world

    state is uniformly accessible, other types of

    knowledge in Homer is restricted

    3.8 Efficiency of Knowledge Access

    As the knowledge bases of these architectures

    grow increasingly large, an efficient way of

    accessing specific knowledge in the knowledge

    bases becomes increasingly important. While this

    is an issue for almost all architectures, MAX has

    good and bad aspects for efficient knowledgeaccess. The good aspects are that all knowledge is

    homogeneously represented and uniformly

    accessed using frames. The bad aspects are that all

    knowledge is declaratively stored.. In the case of

    SOAR , with a very large knowledge base

    containing perhaps a million or more productions,

    accessing the knowledge in an efficient manner is

    an important issue for the architecture. During the

    Elaboration Phase in Soar, all the productions in

    the memory that apply are fired in parallel. This

    parallel operation allows the architecture to access

    all the relevant knowledge in the long-term

    memory in an efficient, tractable way.3.9 Knowledge Consistency

    Knowledge consistency is the property that a

    knowledge database contain no contradictions. It

    is extremely important for knowledge

    representations that may only either assert or deny

    statements, with no measure of partial belief. One

    such system is first order predicate calculus.

    Because all statements may be either true or false,

    it may be possible to store only part of the

    statements , from which all true statements (or all

    false statements) may be derived. Statements that

    can't be derived from the basis set are assumed

    false (or true). This is the closed-world

    assumption. The primary advantage of knowledge

    consistency is ability to store less statements.

    An architecture that tolerates knowledge

    inconsistency generally treats its knowledge base

    as a set of competing hypotheses, or as a set of

    statements that it has varying amounts of

    confidence in. Often there is a numerical measure

    of belief. The flexibility in representation and

    flexibility in learning and reasoning are some of

    the advantages .

    Prodigy is a consistent knowledge

    architecture while MAX is a inconsistent

    knowledge architecture .

    3.10 Homogeneous (Uniform) Knowledge

    Representation

    Architectures with global access to

    knowledge may store it in a uniform format in a

    central database, or may have it in a non-uniformformat in a distributed fashion. The uniform

    method is often employed in systems with a

    general knowledge representation scheme like

    frames or first order predicate calculus. The non-

    uniform method is often used in loosely-coupled

    architectures for storing special knowledge not for

    general use . Access all knowledge easily ,

    interfaces not required to be changed by addition

    or change of modules and easy to modify module

    to accommodate new knowledge are the some of

    the advantages of knowledge uniformity . The

    design limitation , inefficiency to work on a few

    modules and wastage of memory are some of theadvantages .

    In Prodigy, all types of rules- inference rules,

    operators, and control rules, are represented

    uniformly. This allows the control rules to apply

    to both inferences and operators, guiding the

    application of rules. This representation allows

    Prodigy to glass box approach.

    3.11 Heterogeneous Knowledge Representation

    The non-uniform method is often used in

    loosely-coupled architectures for storing speciality

    knowledge not for general use. A heterogeneous

    knowledge representation is based on associatinga finite automaton with first order logic. A number

    of knowledge representation problems raised by

    the electromyography test features are examined

    in this study and the expert system architecture

    allowing such a knowledge modeling are laid out.

    [6] .Entropy Reduction Engine uses

    Heterogeneous Knowledge Representation .

    3.12 No Explicit Representation

    Usually the representation of knowledge

    takes some form. Examples include: first-order

  • 8/7/2019 Agent and the Knowledge Representation-6_411

    5/10

    predicate logic, frames, networks, scripts, etc.

    However, knowledge does not have to be

    represented explicitly. The subsumption and

    ATLANTIS agents rely completely on the world

    for information. There is no explicit knowledge

    representation -- no global representation of state

    nor any underlying symbol level. "The world is its

    own best model"[7].

    4 SPECIFIC EXAMPLES OF

    REPRESENTATIONS AND MEMORY

    STRUCTURES

    4.1 Associative Memory

    Associative memory refers to a memory

    organization in which the memory is accessed by

    its content . Thus, reference clues are "associated"

    with actual memory contents until a desirable

    match is found. Production systems are obvious

    examples of systems that employ such a memory.Associative memory stands as the most likely

    model for cognitive memories, as well. Humans

    retrieve information best when it can be linked to

    other related information. This linking is fast,

    direct and labyrinthian in the sense that the

    memory map is many-to-many and homomorphic

    [8].

    Associative Memory is a primary property of the

    Soar .

    4.2 Episodic Knowledge

    Two particular types of knowledge --

    procedural and declarative -- have been used

    extensively in the design and development of

    cognitive architectures. However, neither of these

    types of knowledge characterize the knowledge

    that humans use to remember events. Such

    remembrances are called episodic knowledge.

    Since this knowledge is, by definition,

    experiential, it must learned by the agent rather

    than pre-encoded (of course, it is possible to

    conceive of episodic knowledge being pre-

    encoded but, once running, additions to the

    episodic memory would then be learned). This

    knowledge can confer the capability to perform

    protracted tasks or to answer queries about

    temporal relationships and to utilize temporal

    relationships.Episodic knowledge is a key feature of the Homer

    architecture.

    4.3 First-order Logic Representation

    Many of the architectures analyzed build

    upon a substrate of First Order Predicate Calculus.

    This is a very descriptive declarative

    representation with a well founded method of

    deriving new knowledge from a database. Its

    flexibility makes it a good choice when more than

    one module may add to or utilize a common

    database (c.f. Prodigy). Unfortunately, this

    flexibility has limitations. First-order predicate

    logic is composed of statements that are assumed

    to be true. The statements are composed of atoms

    ,predicates, two substatements joined by a

    conjunction, disjunction, or implication, a negated

    substatement, and a statement with an existential

    or universal quantifier (in this case, atoms in the

    statements can be replaced by variables in the

    quantifier). [1]

    4.4 Strips-like Operator Representation

    STRIPS, or the Stanford Research Institute

    Problem Solver, was proposed by Fikes and

    Nilsson in 1971[1] and included a representation

    for operators that was intended to solve the frame

    problem. STRIPS uses well-formed formulas of

    the first-order predicate calculus and specifies

    operators by a precondition list, an add-list and a

    delete-list. The preconditions must be satisfied by

    the current state before an operator is applied. Theeffects of the operator are given by the add and

    delete lists. The add-list adds new, instantiated

    well-formed formulas (or wffs, logical descriptions

    of the world) to the current state. The delete-list

    removes wffs from the current state.

    Although STRIPS did resolve some of the

    issues related to the frame problem, it (and all

    systems that use a STRIPS-like representation)

    suffers from a requirement for explicitness -- all

    actions (including secondary effects) must be

    included in the model of the operator. In complex

    worlds, this is often impossible [10].

    4.5 Frame-Like RepresentationMarvin Minsky in the book on computer

    vision proposed frames as means of representing

    common-sense knowledge. Here knowledge is

    organized in to small packets called frames. The

    contents of the frame are certain slots which have

    values.

    To example the concept of frames, try to

    recall the structure of our computer centre.

    Various equipments like computer system, dump

    terminals, printers, air-conditions, un-interrupted

    power supplies etc. are placed at some position in

    the computer centre. By this time we have formed

    an image of computer centre in mind and it iseasy for us to identify without going there where

    things are kept. Suppose we visit a new computer

    center. Our mind now considers a series of

    permutation and combinations . The frame of the

    computer centre in mind is oriented towards the

    present situation with some slots being added (for

    new objects) and some slots being removed. For

    every equipment /object, we form a frame.

    A frame can be defined as a data structure that

    has slots for various objects and a collection of

  • 8/7/2019 Agent and the Knowledge Representation-6_411

    6/10

    frames consists of various objects and a collection

    of frames consists of expectations for a give

    situation. A frame structure provides facilities for

    describing objects, facts about situations,

    procedures in what to do when a situation is

    encountered. Because of there facilities a fames

    provides, frames are used to represent the two

    types of knowledge, namely declarative factual

    and procedural. Like a semantic network, one of

    the chief properties of frames is that they provide

    a natural structure for inheritance. ISA Links

    connect classes to larger parent classes and

    properties of the subclasses may be determined at

    both the level of the class itself and from parent

    classes.

    In addition, the values of a particular attribute

    need not necessarily be filled with a value but may

    also indicate a procedure to run to obtain a value.

    This is known as an attached procedure. Attached

    procedures are especially useful when there is a

    high cost associated with computing a particularvalue, when the value changes with time or when

    the expected access frequency is low. Instead of

    computing the value for each instance, the values

    are computed only when needed. However, this

    computation is run during execution (rather than

    during the establishment of the frame network)

    and may be costly.

    THEO , Homer , Meta Reasoning

    Architecture extensively use frames for

    representation .

    In the Table1A and Table 1 B , Rows indicate the

    Memory , Knowledge and knowledgerepresentation and the column indicates the

    particular architecture . Y in the cell means that

    the architecture corresponding to column has the

    type of memory or knowledge and knowledge

    representation .

    Table 1A

    Memory,Knowledge

    &

    Representation

    Subsumption

    ATLANTIS

    THEO

    PRODIGY

    ICARUS

    ADAPTIVEINTELLIGENT

    SYSTEMS

    Forward &Back-ward Y Y

    Chaining

    Impasse-drivenControl Y YSerial

    Processing Y YParrellelProcessing Y Y

    AsynchronousProcessing Y YInterruptibleProcessing Y YOpen-Loopprocessing Y

    Closed -LoopProcessing Y

    HierarchicalOrganization Y Y Y YModularOrganization Y Y Y Y

    SymbolicWorldModel Y Y Y Y Y

    Size of theKnowledge -Base Y YGlass Box

    approach Y YBlack BoxApproach Y Y

    DeclarativeRepresentation Y Y YProceduralRepresentationGlobalRepresentation Y Y Y Y

    Uniform

    Access toKnowledge Y YKnowledgeConsistency Y YHomogenous ( Uniform)KnowledgeRepresentation Y Y Y Y

  • 8/7/2019 Agent and the Knowledge Representation-6_411

    7/10

    HeterogeneousKnowledgeRepresentationNo-Explicit

    Representation Y

    AssociateMemory

    EpisodicKnowledgeMeta-Knowledge YFirst-OrderLogicRepresentation Y

    STRIPS-LIKEOPERATORREPRESENTATION Y

    Frame -LikeRepresentation YNetworkrepresentation Y Y

    Table 1B

    Memory,Knowledge&

    Representation

    METAREASONING

    ARCHITECTURE

    HOMER

    SOAR

    TETON

    RALPH-MEA

    ENGINE

    Forward &Back-wardChaining YImpasse-drivenControl Y Y YSerialProcessing Y Y Y Y

    ParallelProcessing Y Y Y

    AsynchronousProcessing Y

    InterruptibleProcessing Y Y

    Open-Loopprocessing Y

    Closed -LoopProcessingHierarchicalOrganization Y Y

    ModularOrganization Y Y Y YSymbolicWorld Model Y Y Y Y YSize of theKnowledge -Base YGlass Boxapproach Y Y

    Black BoxApproach Y Y YDeclarativeRepresentation Y Y YProceduralRepresentation Y Y

    GlobalRepresentation Y Y Y Y Y YUniformAccess to

    Knowledge Y YKnowledgeConsistency Y

    Homogenous( Uniform)KnowledgeRepresentation Y

    HetrogenousKnowledgeRepresentation Y Y YNo-Explicit

    RepresentationAssociateMemory YEpisodicKnowledge Y

    Meta-Knowledge Y Y Y YFirst-OrderLogicRepresentation

  • 8/7/2019 Agent and the Knowledge Representation-6_411

    8/10

    STRIPS-LIKEOPERATORREPRESENTATION Y YFrame - LikeRepresentatio

    n Y YNetworkrepresentation

    Table 2

    Sl

    N

    o

    Architect

    ure

    Features

    1 Subsumpt

    ion

    Complicated Intelligent

    behavior

    Into

    Simple behavior modules

    Organized into

    layers

    2 THEO Plan-Then-Complie

    By this

    Integrating learning ,

    planning and knowledge

    representation

    3 ICARUS Specific representation of

    long term memory .It uses

    3 independent

    asynchronous modules

    responsible for

    1.Perception

    2.Planning3.Effecting

    4 PRODIG

    Y

    Storing the knowledge in a

    form of first order predicate

    logic (FOPL) called

    Prodigy Description

    language (PDL) .

    Has a modular architecture

    that stores the knowledge

    symbolically.

    5 ATLANT

    IS

    Integrating planning and

    reacting in a heterogeneous

    asynchronous architecture

    for mobile agents .

    It consists of 3 layers:

    1.Control layer

    2.sequencing layer

    3.deliberative layer

    6 Adaptive

    Intelligen

    t System

    (AIS)

    To reason about and

    interact with other dynamic

    entities in real time .

    --- problem solving

    techniques

    --- when encountering un-

    expected situation , decides

    whether to and how to

    respond .

    --- focus attention on the

    most critical aspects of

    current situation .

    --- operating continuously

    without rebooting .

    --- able to coordinate with

    external agent .

    (more or less similar to

    human being )

    7 Meta

    Reasonin

    g (MAX)

    Many ideas in MAX may

    traced to Prodigy.

    --- rule-based forward

    chaining engine that

    operates on productions .

    --- is designed to support to

    modular agents.

    --- they are used to

    respond to a dynamicenvironment in a timely

    manner .

    --- modules are categorized

    in to Behavior and monitor

    .

    --- Some of the modules

    are:

    1.attention focusing

    2.multiple problem solving

    strategies

    3.execution monitoring

    4.goal-directed exploration5.explantion-based learning

    6.process interruption

    7.intelligent resumption

    8 HOMER --- Is not designed for

    general intelligence .

    --- underlying philosophy is

    to synthesize several key

    areas of AI to form one

    complete system . (like

    planning, learning , natural

    language understanding ,

    robotic navigation ) .

    HOMER answers questionsposed by users and carries

    out instructions given by

    users .

    --- is a modular structure.

    It consists of :

    1.memory

    2.planner

    3.natural language

    interpreter and generator .

    4.reflective processes

  • 8/7/2019 Agent and the Knowledge Representation-6_411

    9/10

    5.plan executer

    9 SOAR --- originally known as

    STATE OPERATOR AND

    RESULT .

    --- main goal is that full

    range of capabilities to behandled by an intelligent

    agent from highly routines

    to extremely difficult open-

    problems

    --- the underlying SOAR

    architecture is the view that

    symbolic system is

    necessary and sufficient

    condition for general

    intelligence . This is known

    as Physical Symbolic

    system Hypothesis (PSSH)

    --- ultimate aim is to get

    general intelligent agent--- is based on a production

    system ie. It uses explicit

    production rules to govern

    its behaviors .

    10 Teton --- is a problem solver

    --- uses two memory areas

    1.Short-Term

    memory

    2. Long-Term

    memory

    --- like human beings ,interruption are allowed .

    --- it has a feature called

    Execution Cycle which

    always look for what to do

    next .11 RALPH-

    MEA

    --- is a multiple execution

    architecture

    --- like human being ,

    selecting best one from the

    environment

    --- RALPH MEA uses

    Execution Architecture

    (EA) to select from onestate to best one .

    --- it uses the following :

    1.Condition action

    2.Action utility

    3.Goal based

    4.Decision Theoretic

    12 Entropy

    Reductio

    n Engine

    (ERE)

    --- focuses on problems

    that require planning ,

    scheduling and control

    --- uses many different

    problem solving methods

    such as :

    1.problem reduction

    2.temporal projection

    3.rule-based execution

    5 CONCLUSION

    The problem of AI is to describe and build agents

    that receive percepts from the environment and

    perform actions, and each such agent is

    implemented by a function that maps percepts to

    actions. It explains the role of learning as

    extending the reach of the designer into unknown

    environments, and shows how it constrains agentdesign, favoring explicit knowledge representation

    and reasoning . It analyzes basic techniques for

    addressing complexity . The idea is to Integrate

    state-of-the art AI techniques into intelligent agent

    designs, using examples from twelve agents to

    full knowledge-based agents with natural

    language capabilities and so on . This leads to the

    study of Multi-Agent systems and its applications

    . In depth analysis of various Agent architectures

    and their capabilities is to build a Multi Agent

    System that will be suitable for our future work on

    Supply Chain Management

    REFERENCE

    [1] Elavine Rich , Kevin Knight and Shivshankar

    , Artificial Intelligence by The Mcgraw Hill

    Publishing Company limited .

    [2] Kurt Vanlehn , Rule-Learning Events in the

    acquisition of a complex skill ,

    Journal article by, Journal of the Learning Science

    ,vol 8 , 1999

    [3] Philip E. Agre , Hierarchichy and History in

    Simons Architecture of Complexity Journal of

    Learning Science 12(3) , 2003

    [4] Newell a , Unified theories of Cognition ,

    Harvad University Press , Cambridge ,

    Massachusetts , 1990

    [5] Levesque H , Brachman R , A fundamental

    trade off in Knowledge representation and

    reasoning , 1998

    [6] Vincent Rialle , Annick Vila , Yves Besnard ,

    Artificial Intelligence in Medicine 1991

    [7] Vincent C.Muller , Is there a future for AI

    without representation , springer Verlag , 2007

    [8] Pat Langley , Dongkyu Choi , Seth Rogers ,

    Interleaving Learning Laboratory Centre for the

  • 8/7/2019 Agent and the Knowledge Representation-6_411

    10/10

    study and languages and information , Stanford

    University , Stanford .

    [9] A.S Maida , A Uniform architecture for rule-

    based meta-reasoning and representation :- case

    study . , IEEE Computer Society Press.

    [10] Artificial Intelligence: Modern Approach by

    Stuart J. Russell , Peter Norvig Prentice HallSeries in Artificial Intelligence.