INTELLIGENT AGENTS

1
INTELLIGENT AGENTS

• IS4185 Session 15
• Prof. Mark Nissen

2
Agenda

• Agent Motivation Functionalities
• Agent Background Definitions
• Conceptual Framework
• Agent Examples
• Inter-agent Cooperation Competition
• Acquisition Process Redesign
• Intelligent Supply Chain Agents
• Process Vision Challenges (demo)

3
Agent Motivation

• People are limited, overworked
• Machines are deterministic, incessant
• Knowledge work factors
• Data collected by large enterprises 2x/year
• KWs can analyze only 5 of this data
• KWs spend time
• 60 trying to discover data patterns
• 20 interpreting patterns
• 10 acting on patterns interpretations
• Save time money, increase efficacy

4
Agent Functionalities

• Overcome information overload
• Automate mundane info tasks
• Persistently vigilantly serve master
• Interface with complex systems
• Support/make decisions
• Cognitive prostheses
• Design keys
• Easy to build small, limited KBS
• Link together in cooperative federation

5
Agent Background

• Early concepts
• Bushs memex - 40s
• McCarthys AdviceTaker - 50s
• DAI research - late 70s
• Distributed, expert systems technology
• Centralized design and architecture
• Remote action research - mid 90s
• From reasoning to doing
• Limited intelligence capability
• Decentralized, heterogeneous agents

6
Collaboration
Parallel processing
ISCA
Expert systems
Intelligence
Remote programming
Mobility
7
Agent Definitions

• Disagreement on agent definition
• Classify by attributes (intel, mobility)
• Classify by intention (transact, represent)
• Useful abstraction (module, object)
• Many diverse agent examples
• Information filtering
• Information retrieval
• Advisory
• Performative

8
Agent Examples

• Information filtering - I search
• User e-mail preferences (Maes, Malone)
• NetNews preferences (Sycara Zeng)
• FAQs search filtering (Whitehead)
• Arbitrary text (Verity)
• Information retrieval - I search
• Compact disks (Krulwich, BargainFinder)
• Computer equipment (uVision)
• Advertising rates (PriceWatch)
• Insurance services (Insurance)

9
Agent Examples

• Information retrieval (cont)
• Web indexing robots (Etzioni Weld)
• Web report writing (Amulet)
• Web publishing (InterAp)
• Assisted browsing (Burke)
• Tech info delivery (Bradshaw)
• Info gathering (Knobloch Ambite)
• Books (BargainBot) http//www.ece.curtin.edu.au/s
  aounb/bargainbot/

10
Agent Examples

• Advisory - I recommendations
• Music recommendations (Maes, Firefly)
• Electronic Concierge (Etzioni Weld)
• Campus visit host (Zeng Sycara)
• Manufacturing plans (Maturana Norrie)
• Strategic planning (Pinson)
• Software project coordination (Johar)
• Interface assistance (Ball)
• Reconnaissance (Bui)
• Portfolio management (Sycara)

11
Agent Examples

• Performative - I behavior
• Agent2agent markets (Chavez Maes)
• Negotiation (Bui)
• Scheduling (Sen, Walsh)
• Cooperative learning (Boy)
• Digital services (Mullen Wellman)
• Software purchasing (Mehra Nissen)

12
Inter-Agent Cooperation Competition
Agents Orientation Design Cooperative C
ompetitive Centralized Predetermined n/a
Distributed Trust-based Negotiation Job
specialization Game-theoretic
13
Inter-Agent Cooperation Competition

• Inter-agent protocols
• KIF - Declarative - predicate calculus
• KQML - Standard communication types
• IIOP - Net inter-ORB protocol
• Negotiation - multiple bids quotes
• Game-theoretic reasoning

Unknown/Unfriendly Agent Friendly
Agent Cooperative Competitive
Cooperative 10 -5 Competitive 5 -10
14
Acquisition Domain

• DoD term, mil/com/gov/org application
• Procurement planning, contracting,
  program/project management logistics
• Managing relationships materiel
• Attaining strategic importance
• Limit fleet/battlefield speed mobility
• Supply-chain integration IT, BPR, KM
• Virtual orgs, competition vs. cooperation
• Maintain dynamic inter-org networks

15
Acquisition Pathologies

• Low level procurement/supply orgs
• Focus on products, not relationships
• Bureaucratic functions seldom integrated
• Manual, paper-based, labor-intensive
• Slow, expensive, rigid, unadaptable
• Complex laws, policies, international regs
• Dynamic, time-sensitive markets
• Unresponsive to users customers
• Process unable to meet user/customer requirements
  market dynamics

16
Acquisition Process Redesign

• Redesign S/W purchasing process
• Baseline transformation
• Model baseline process
• Measure diagnose pathologies
• Recommend redesign transformations
• Simulation, selection implementation
• Integrated supply-chain process
• NPS purchasing
• Gensym order fulfillment

17
Baseline Process
18
Baseline Digraph (partial)
ID Mkt PR Verify
Rsch Issue Prep rqmts survey
form form sources RFP quote
...
User User User SupO SupO
SupO Ktr
HO
HO
VCX
... NPS NPS NPS NPS
NPS NPS Ktr1... Web
DBMS ... Doc1
Doc2 Doc3 Doc4
Doc5 Doc6 ... P P
P P P
P etc.
19
Baseline Measurements
20
Agent Redesign Opportunities

• Intelligent, autonomous software agents
• Radical redesign acquisition reform
• Leverage AI, IT network advances (EC)
• Reduce cost cycle time
• Improve quality responsiveness
• Virtual re-intermediation of procurement
• Directly connect users contractors
• Follow all laws, regs, policies, rules, etc.
• Maintain deep product industry K
• Monitor markets, suppliers, prices, laws

21
Intelligent Acquisition Agents

• Multi agent system - proof-of-concept
• 3 agent classes (user, supply, vendor)
• Distributed hosts processing
• Agent coordination along supply chain
• ADE - shell-like tool
• Grafcets --gt G2 environment or Java
• Distributed agent messaging classes
• Behaviors implemented via methods
• Autonomous, mobile, intelligent, cooperative,
  multi-threaded agents

22
ISCA Architecture
ADE - Agent Development Environment - Within
(networked) G2 systems (objects methods) -
Heterogeneous JVM platforms (JavaBeans) -
Multi-threaded agents, inter-agent messaging
23
ADE Elements

• Agent - compound, mobile object
• Activity - agent behaviors (methods)
• Message - structured classes
• Host - machine environment
• Environment - federation environment
• Simulation - emulate physical envir
• Bridges - PLCs, networks, DBs, etc.

24
Grafcet Specification

• Grafcet - international standard
• Unambiguous behaviors conditions
• Extension of Petri Nets
• Steps, transitions, tokens conditions

Waiting Enabled Fired
25
User Grafcet
26
Supply Grafcet
27
Vendor Grafcet
28
Intelligent Mall

• Mall metaphor
• More complex, robust than supply chain
• More demanding agent behaviors
• Virtual shops and shoppers
• Unique agents, nearly unlimited in number
• Represent sellers and buyers
• Autonomous, long-lived commerce
• Procurement domain obvious
• NPRDC officer billets market

29
Grafcet behavior flows
Objects, methods messages
Virtual mall/supply chain Intelligent
procurement agents
30
Web-based I/O
Suppliers register product data
Users specify requirements
Item, qty, budget, date, etc.
Items, specs, price, delivery, etc.
31
Regulatory environment
Agents locate select suppliers, virtually,
arrange terms
32
Agents complete transactions, return with
goods/contracts
33
Process Vision Challenges

• Direct acquisition V re-intermediation
• Paperless people-free process
• Agents manage routine work (80/20)
• People manage agents, relationships
• Some challenges
• Education vs. training, people agents
• Human/machine work sharing
• Acquisition agent knowledge engineering
• Agent designs, experiments, security, trust
• I-Mall demo