1. Chapter 16: Multiagent Systems Service-Oriented Computing: Semantics, Processes, Agents – Munindar P. Singh and Michael N. Huhns, Wiley, 2005

2. Chapter 162Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns Highlights of this Chapter Applicability in Service-Based Systems Multiagent Architecture Agent Types Lifecycle Management Consistency Maintenance Modeling other Agents Cognitive Concepts

3. Chapter 163Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns Basic Problems of MAS Distributing control among agents Describing, decomposing, distributing tasks Either the designer does it Or the agents negotiate it autonomously Designing agents to support the above Interacting and communicating Negotiating Special case in cooperative, homogeneous setting: maintaining consistency

4. Chapter 164Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns (de facto) Standard Agent Types Design-time (static) breakdown of functionality and tasks

5. Chapter 165Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns Brokerage Service Viewed as an Agent Type Cooperates with a Directory Service Accepts requests from agents to recruit one or more agents who can provide a service Uses knowledge about the requirements and capabilities of registered agents to Identify appropriate agents for an interaction Negotiate with selected agents Potentially learn about the properties of the responses Example: Brokerage determines that advertised results from agent X are incomplete and seeks a substitute for X

6. Chapter 166Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns FIPA Agent Management System Foundation for Intelligent and Physical Agents (now in IEEE) Good Architecture (notions of agents and interaction) Wrong Mentalist focus Protocol models are overconstrained

7. Chapter 167Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns Agent Management System: 2 Analogous to a JEE Container Handles the creation, registration, location, communication, migration, and retirement of agents White pages, e.g., agent location and naming (from AMS) Agent identifiers support social names, transport addresses, name resolution services Yellow pages, e.g., service location and registration services (from DF) Agent message transport services

8. Chapter 168Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns Agent Frameworks JADE, a popular FIPA-compliant agent framework for multiagent systems: http://jade.tilab.com/ (FIPA-OS and Zeus having died) Jadex: JADE plus BDI constructs JaCoMo: Combines three programming approaches Jason: BDI constructs Cartago: Environment artifacts Moise: Organizations

9. Chapter 169Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns Consistency Maintenance across Services A truth maintenance system (TMS) helps maintain consistency Performs a form of propositional deduction Maintains justifications and explains the results of its deductions Updates beliefs incrementally when premises change TMSs help us Deal with atomicity Maintain modular models

10. Chapter 1610Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns Architecture of TMS-Based Agent Problem solver: decides on actions TMS: maintains a network of beliefs Justifications of a belief based on others help make corrections to reasoning Problem Solver TMS justifications beliefs

11. Chapter 1611Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns Knowledge Base Integrity Stability: believe everything justified validly; disbelieve everything else Well-Foundedness: no circular beliefs Logical consistency: no logical contradictions Completeness: find a consistent state if one exists, or report failure Problems arise when knowledge is distributed

12. Chapter 1612Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns Distributed TMS Each agent has a justification-based TMS Each datum can have status OUT (not believed) IN: valid local justification (believed) EXTERNAL: must be IN for some agent When a problem solver adds or removes a justification, the DTMS checks if some datum is affected Unlabels data based on the changed datum Relabels all unlabeled shared data (in one or more iterations) Notifies other agents who have an interest in any datum, i.e., those with whom the datum is shared

13. Chapter 1613Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns Degrees of Logical Consistency Inconsistency: an agent is internally inconsistent All bets are off with such an agent Local Consistency: all agents are individually consistent Totally disconnected agents can’t interact effectively Global Consistency: union of KBs is consistent Total integration is not viable in open settings [DTMS] Local-and-Shared Consistency: agents are locally consistent and agree about any data they might share Captures essential interdependence

14. Chapter 1614Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns Cooperative Service: 1 Client f3: afford(xcorp) IN r3: buy(X) :- query(Broker recommend(X)), afford(X) IN Broker f1: afford(xcorp) OUT f2: cash-rich(xcorp) IN r2: recommend(X) :- takeover-bid(X) IN r1: takeover-bid(X) :- cash-rich(X) IN ? recommend(?X)

15. Chapter 1615Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns Cooperative Service: 2 Client f3: afford(xcorp) IN r3: buy(X) :- query(Broker recommend(X)), afford(X) IN Broker f1: afford(xcorp) OUT f2: cash-rich(xcorp) IN r1: recommend(X) :- takeover-bid(X) IN r2: takeover-bid(X) :- cash-rich(X) IN f3: recommend(xcorp) IN Shared with: Client; Justification: (f2 r1 r2) recommend(XCorp)

16. Chapter 1616Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns Cooperative Service: 3 Client f3: afford(xcorp) IN r3: buy(X) :- query(Broker recommend(X)), afford(X) IN f4: recommend(xcorp) EXTERNAL Shared with: Broker; Justification: ( ) f5: buy(xcorp) IN Justification: (f3 f4 r3) Broker f1: afford(xcorp) OUT f2: cash-rich(xcorp) IN r1: recommend(X) :- takeover-bid(X) IN r2: takeover-bid(X) :- cash-rich(X) IN f3: recommend(xcorp) IN Shared with: Client; Justification: (f2 r1 r2)

17. Chapter 1617Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns Cooperative Service: 4 Client f3: afford(xcorp) IN r3: buy(X) :- query(Broker recommend(X)), afford(X) IN f4: recommend(xcorp) EXTERNAL Shared with: Broker; Justification: ( ) f5: buy(xcorp) IN Justification: (f3 f4 r3) Broker f1: afford(xcorp) OUT f2: cash-rich(xcorp) IN --> OUT r1: recommend(X) :- takeover-bid(X) IN r2: takeover-bid(X) :- cash-rich(X) IN f3: recommend(xcorp) IN --> OUT Shared with: Client; Justification: (f2 r1 r2) relabel recommend(XCorp)

18. Chapter 1618Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns Cooperative Service: 5 Client f3: afford(xcorp) IN r3: buy(X) :- query(Broker recommend(X)), afford(X) IN f4: recommend(xcorp) OUT Shared with: Broker; Justification: ( ) f5: buy(xcorp) OUT Justification: (f3 f4 r3) Broker f1: afford(xcorp) OUT f2: cash-rich(xcorp) OUT r1: recommend(X) :- takeover-bid(X) IN r2: takeover-bid(X) :- cash-rich(X) IN f3: recommend(xcorp) OUT Shared with: Client; Justification: (f2 r1 r2)

19. DTMS Scope Ensures consistent state with respect to shared data Considers one state of the world Even though agents may learn or unlearn data about the same state Not suitable for dealing with a changing world Cannot deal with real-world actions that an agent cannot undo Chapter 1619Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns

20. Chapter 1620Service-Oriented Computing: Semantics, Processes, Agents - Munindar Singh and Michael Huhns Chapter 16 Summary Study multiagent systems because interactions among agents make them interesting Communication among agents is key, although markets (later chapter) only support implicit communication through prices Programming environments support agent interactions Consistency maintenance is a major challenge Agents must model agents; simple techniques are often adequate; more subtle techniques can require extensive reasoning power