Presentation is loading. Please wait.

Presentation is loading. Please wait.

Agent Factory : A Prototyping Environment for Agents Gregory O’Hare CS 4019 Multi-Agent Systems Lecture Material, University College Dublin.

Similar presentations


Presentation on theme: "Agent Factory : A Prototyping Environment for Agents Gregory O’Hare CS 4019 Multi-Agent Systems Lecture Material, University College Dublin."— Presentation transcript:

1 Agent Factory : A Prototyping Environment for Agents Gregory O’Hare CS 4019 Multi-Agent Systems Lecture Material, University College Dublin

2 Outline What is the difference between Objects and Agents? What is Agent Factory? Visual Design within Agent Factory Case Study: WAY System Agent Factory in Context Conclusions

3 Objects versus Agents Numerous Authors & Researchers have debated the differences and similarities between Objects and Agents; Agents are autonomous. An object does not exhibit control over its behaviour; methods are made available for invokation as and when desired; Agents do not invoke methods but make requests; Objects have nothing to say about differing deductive models like reactive, pro-active or social abilities Agents are each considered to have their own thread of control. In standard object systems there is merely one thread; However some concurrent object based programming languages starting to emerge;

4 Active Objects Active Objects blur the distinction; Active objects have their own thread of control and can in some senses be considered autonomous; They exhibit some behaviours without actually being operated upon;

5 What is Agent Factory? Agent Factory is… “a cohesive framework that supports a structured approach to the development and deployment of multi-agent systems.” AF-APL is augmented with: AF-APL Run-Time Environment Development Environment Development Methodology

6 AF-APL AF-APL is a declarative language. Designed to simplify the implementation of complex agent behaviours. Underpinned by a formal agent theory. Founded upon an analysis of how agents commit to actions. Formalised as a multi-modal temporal logic. Specifically, Agent Factory agents employ: Beliefs Commitments Commitment Rules

7 AF-APL Control Algorithm World Belief Update Query Perceptors Actuators Beliefs Commitments Commitment Management Commitment Rules Migrate? 1 Perceive 2 Deliberate 3 Act

8 AF-APL AF-APL Programs define: Actuators Perceptors Modules Commitment Rules Initial Mental State

9 The Run-Time Environment Deals with Interoperability. FIPA compliant Design It is comprised from one or more Agent Platforms (APs) Each AP combines: Support for agent migration White and Yellow Pages Agents Agent Management Services Initial Agent Community

10 The Agent Design Tool Supports reuse of AF-APL code through inheritance. Implements filtered views of agent classes. Facilitates “publication” of agent designs.

11 Agent Viewer Tool Supports the visualisation, configuration, and execution of agents Provides a filtered view of the agent through a set of views

12 The Development Methodology Defines a structured approach to the coding and testing phases. Places a partial ordering on the use of the tool set. Identifies key deliverables to be produced.

13 Visual Design with Agent Factory Provision of graphical tools to support the design of agents. Initial focus upon support for the modelling of agent interaction protocols. Provides scope for the utilization of various design notations. Supports users in the fabrication of semantically correct protocols. Assist users in the generation of agent code based upon published protocols. Initial prototype supports Agent UML notation.

14 Agent Factory Protocol Tool Interface Semantic Checker User Updates Model Update of Model Inform of failureInform of success

15 Agent Factory Protocol Tool

16 M-Commerce. Agent Factory is being used to deliver context sensitive shopping for mobile users. EasiShop Robotics. Agent Factory was employed as part of a hybrid agent architecture that is being used in ongoing research to investigate social robotics. IMPACT, MARC Speech Processing. Agent Factory is being used to process human speech using Phonotactics. SAID Current Applications Ubiquitous Computing. Agent Factory has been employed to deliver a number of location aware services for hand-held devices. Gulliver’s Genie, WAY System, E=mc2, and AD-ME

17 The Goal of the WAY Support the mobile citizen in the location, tracking and rendezvousing with a variety of moving entities. Where are you? I’m walking down Grafton Street Well, I’m outside Trinity College Grand, I’ll see you in a minute Allow the mobiles themselves to deal with these requests

18 Technologies for the WAY Personal Digital Assistants (PDAs) Personal Digital Assistant (IPAQ 3600 - 3700) PCM/CIA Sleeve Localisation Technology including Global Positioning Systems (GPS) GPS card / Garmin 2 GPS receiver Communication over Wireless Networks Nokia Card Phone 2 GSM 900/1800 Mobile, Lightweight, Belief Desire Intention (BDI), Agents Agent Factory

19 Implementing the WAY Each user is given their own WAY agent Maintains information about the users current location and activities By preference, WAY Agents reside on the users mobile device When on the device, WAY Agents have access to the users position through GPS Also, they can interact with the user through a graphical interface However, if the device is not accessible, the agents migrate to the WAY Server While there, they remember the users last known coordinates. WAY Agents communicate regular updates of their users position to interested parties But only if the user permits it!

20 At the end of the WAY… Zoomed InZoomed Out

21 Agent Factory in Context I A number of other Agent Development Tools exist: LEAP (LEAP Consortium). Integration of JADE and ZEUS that is compliant with J2ME. JADE (TILAB). FIPA-compliant Java API that supports the fabrication of reactive agents. ZEUS (BT Labs). A graphical toolkit for creating deliberative agent designs, which when completed, are compiled into Java code, customised and finally, executed. JACK (Agent-Oriented Software). Extends Java with agent-based concepts. JACK code is compiled into Java code and executed. FIPA-OS (Emorphia). The first FIPA-compliant agent platform. Similar to JADE.

22 Agent Factory in Context II Table 1. A Comparison of Agent Prototyping Environments LEAPJACKZEUSJADEFIPA-OS AF BDI √ √ Mobility √√ √√ White Pages √√ √√√ Yellow Pages √√ √√√ FIPA Compliance √√ √√√ Fabrication ModeDesignInstanceDesignInstanceDesign Inheritance √ √ √√ ConstructionGraphical None VisualizationGraphical NoneGraphicalNone Integrated Methodology √√ √ √ √

23 Summary Agent Factory: Is a cohesive framework for the development and deployment of agent-oriented applications on J2ME-compliant devices. Offers a visually intuitive development environment that: Promotes design reuse. Has strong links with Agent UML. Delivers a rich set of tools to support the development and debugging of agent designs. Provides a rich coupling between an Agent Development Methodology and the Agent Factory toolset. Supports mobile agents.

24 And, finally… An evaluation version of the Agent Factory Framework is available for download from: http://www.agentfactory.com/

25 Agent Factory Physical Architecture Component Design Library Agent Designs Perceptors Actuators Modules MTS MS agentfactory.ucd.ie krytan.ucd.ie Development Environment lister.ucd.ie Development Environment holly.ucd.ie Agent Platform pda1.ucd.ie Agent Platform Published components Deployed components

26 The Actuator Builder Supports the fabrication of Actuators Interface presents a filtered view of the implementation Generates and compiles relevant Java classes


Download ppt "Agent Factory : A Prototyping Environment for Agents Gregory O’Hare CS 4019 Multi-Agent Systems Lecture Material, University College Dublin."

Similar presentations


Ads by Google