Pizza and a Movie

253 views

Published on

A Case Study in Advanced Web Services

Published in: Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
253
On SlideShare
0
From Embeds
0
Number of Embeds
5
Actions
Shares
0
Downloads
3
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide
  • Services whose description language is XML-based (for example, WSDL), whose interfaces are exposed as SOAP interfaces and which are registered in a UDDI directory.
  • The event organizer is an example of service composition, like hotel-transportation-restaurant-transportation-cinema-transportation-hotel. Pervasive computing to retrieve information that is then used for personal profiling in the Agentcities environment.
  • Interoperation between different software systems based on agreed, semantically grounded definitions.
  • core platform and agents: AMS and DF
  • Pizza and a Movie

    1. 1. Pizza and a Movie A Case Study inAdvanced Web Services Dr Luigi Ceccaroni Dr Jonathan Dale Network Agent Research Group Fujitsu Laboratories of America July, 2002
    2. 2. © 2002 Fujitsu Laboratories of AmericaOutline• Introduction• The Evening Organizer• The scenario• Conclusions• Challenges
    3. 3. © 2002 Fujitsu Laboratories of America Introduction• Advanced Web Services (WS++) – Communicate on wide-area networks, such as the Internet. – Have explicit representation of the capabilities offered. – Support semantically grounded communication. – Are reusable in different contexts and situations. – Are published, discovered through agreed registry mechanisms. – Are invoked through interactive and dynamic interfaces.• Agents – Advertise their existence. – Establish communications with other agents. – Describe their function, input and output requirements. – Negotiate an exchange of data and services. – Cooperate in the completion of a task.
    4. 4. © 2002 Fujitsu Laboratories of America The Evening Organizer• Part of Agentcities.RTD composite services• Uses various domain services such as:  Weather  Cinema  Hotel  Restaurant  Taxi• Application for San Francisco• Part of an envisioned pervasive-computing infrastructure (location and context-aware services)• AAP based
    5. 5. © 2002 Fujitsu Laboratories of AmericaThe Evening Organizer User Personal Assistant Agent1 Gather Gather Initiate Evening Gather User Ratings Review Planning Task Preferences Information Information Evening Organizer 7a 2 Approve Compose Refine Itinerary Itinerary 3 Itinerary 4 Modify Seek Gather Itinerary Propose Approval of 5 Itinerary Service 6 Itinerary Information Cancel Itinerary 7b 8 Make Make Itinerary Itinerary Reservations Reservations10a Approve Seek Approval Propose Itinerary of Itinerary with Itinerary with Reservations Reservations 9 Modify Itinerary 10b Book Itinerary Reservations The user can Cancel monitor booking Itinerary execution
    6. 6. © 2002 Fujitsu Laboratories of AmericaInitiating an evening organizationtask A user wants to go to a restaurant to eat a pizza and then go to a cinema to see a Sci-Fi movie.The user is guided, via Web forms, towards a Restaurant and Shows template.(Restaurant (typeOfCuisine:Cuisine = Pizza) Shows = CinemaShow (show:Performance (genre:string = Sci-Fi)))
    7. 7. © 2002 Fujitsu Laboratories of AmericaThe Evening Organizer User Personal Assistant Agent1 Gather Gather Initiate Evening Gather User Ratings Review Planning Task Preferences Information Information Evening Organizer 7a 2 Approve Compose Refine Itinerary Itinerary 3 Itinerary 4 Modify Seek Gather Itinerary Propose Approval of 5 Itinerary Service 6 Itinerary Information Cancel Itinerary 7b 8 Make Make Itinerary Itinerary Reservations Reservations10a Approve Seek Approval Propose Itinerary of Itinerary with Itinerary with Reservations Reservations 9 Modify Itinerary 10b Book Itinerary Reservations The user can Cancel monitor booking Itinerary execution
    8. 8. © 2002 Fujitsu Laboratories of America Composing an itineraryThe personal agent (PA) expands the instantiated template into an instantiated itinerary (Web Services selection and composition). (Taxi (source:string = UserHome) Restaurant … Taxi … Shows … Taxi
    9. 9. © 2002 Fujitsu Laboratories of AmericaThe Evening Organizer User Personal Assistant Agent1 Gather Gather Initiate Evening Gather User Ratings Review Planning Task Preferences Information Information Evening Organizer 7a 2 Approve Compose Refine Itinerary Itinerary 3 Itinerary 4 Modify Seek Gather Itinerary Propose Approval of 5 Itinerary Service 6 Itinerary Information Cancel Itinerary 7b 8 Make Make Itinerary Itinerary Reservations Reservations10a Approve Seek Approval Propose Itinerary of Itinerary with Itinerary with Reservations Reservations 9 Modify Itinerary 10b Book Itinerary Reservations The user can Cancel monitor booking Itinerary execution
    10. 10. © 2002 Fujitsu Laboratories of AmericaRestaurant service services Restaurant QUERY-REF Finder Extended-FIPA-SL Evening Agent INFORM Organizer Extended-FIPA-SL REQUEST KIF INFORM-DONE JTP Knowledge BASE Restaurant Booking SQL SQL database Agent DAML+OIL HTML DAML-S (Web)
    11. 11. Grounding for © 2002 Fujitsu Laboratories of AmericaWeb Services invocation (FinderAgents) (QUERY-REF :sender ( agent-identifier :name PaulOakenfold@ bo.agentcities.net:1099 :addresses (sequence http://bo.agentcities.net:7778 )) :receiver (set ( agent-identifier :name restaurant@ sf.us.agentcities.net :addresses (sequence http://sf.us.agentcities.net:8080/acc ))) :content "( (all ?r1 (and (instance ?r1 daml-s#serviceProfile) (= (value ?r1 restaurant#CuisineType) "restaurant#FrenchCuisine") ) ) )" :language extended-FIPA-SL :ontology (sequence daml-s:http://.../onto1 restaurant:http://...onto3) :protocol fipa-query)
    12. 12. © 2002 Fujitsu Laboratories of AmericaThe Evening Organizer User Personal Assistant Agent1 Gather Gather Initiate Evening Gather User Ratings Review Planning Task Preferences Information Information Evening Organizer 7a 2 Approve Compose Refine Itinerary Itinerary 3 Itinerary 4 Modify Seek Gather Itinerary Propose Approval of 5 Itinerary Service 6 Itinerary Information Cancel Itinerary 7b 8 Make Make Itinerary Itinerary Reservations Reservations10a Approve Seek Approval Propose Itinerary of Itinerary with Itinerary with Reservations Reservations 9 Modify Itinerary 10b Book Itinerary Reservations The user can Cancel monitor booking Itinerary execution
    13. 13. Grounding for © 2002 Fujitsu Laboratories of AmericaWeb Services invocation (BookingAgents) (REQUEST :sender ( agent-identifier :name PaulVanDyk@ bo.agentcities.net:1099 :addresses (sequence http://bo.agentcities.net:7778 )) :receiver (set ( agent-identifier :name beryl@ sf.us.agentcities.net :addresses (sequence http://sf.us.agentcities.net:8080/acc )) ) :content "(action ( ( agent-identifier :name beryl@ sf.us.agentcities.net :addresses (sequence http://sf.us.agentcities.net:8080/acc ))) (process#booking :restaurant#howmany 8 :restaurant#when "2002-06-17-1630" ))" :ontology (sequence process:http://.../onto2 restaurant:http://...onto3) :language extended-FIPA-SL
    14. 14. © 2002 Fujitsu Laboratories of AmericaThe Evening Organizer User Personal Assistant Agent1 Gather Gather Initiate Evening Gather User Ratings Review Planning Task Preferences Information Information Evening Organizer 7a 2 Approve Compose Refine Itinerary Itinerary 3 Itinerary 4 Modify Seek Gather Itinerary Propose Approval of 5 Itinerary Service 6 Itinerary Information Cancel Itinerary 7b 8 Make Make Itinerary Itinerary Reservations Reservations10a Approve Seek Approval Propose Itinerary of Itinerary with Itinerary with Reservations Reservations 9 Modify Itinerary 10b Book Itinerary Reservations The user can Cancel monitor booking Itinerary execution
    15. 15. Final form of the © 2002 Fujitsu Laboratories of Americaitinerary (Taxi (source:string = UserHome) (destination:string = Ft. Mason Ctr., A) Restaurant = Antica Trattoria (seating:RestaurantSeating = Indoor) (typeOfCuisine:Cuisine = Pizza) Taxi (source:string = Ft. Mason Ctr., A) (destination:string = 1345 Bush St) Shows = CinemaShow (show:Performance (title:string = Barbarella) (genre:string = Sci-Fi)) (location:EventHost (name:string = Alliance Franç aise)) Taxi (source:string = 1345 Bush St) (destination:string = UserHome))
    16. 16. © 2002 Fujitsu Laboratories of AmericaThe Evening Organizer User Personal Assistant Agent1 Gather Gather Initiate Evening Gather User Ratings Review Planning Task Preferences Information Information Evening Organizer 7a 2 Approve Compose Refine Itinerary Itinerary 3 Itinerary 4 Modify Seek Gather Itinerary Propose Approval of 5 Itinerary Service 6 Itinerary Information Cancel Itinerary 7b 8 Make Make Itinerary Itinerary Reservations Reservations10a Approve Seek Approval Propose Itinerary of Itinerary with Itinerary with Reservations Reservations 9 Modify Itinerary 10b Book Itinerary Reservations The user can Cancel monitor booking Itinerary execution
    17. 17. © 2002 Fujitsu Laboratories of America Conclusions• EO: a simple process (black-box) expanding to a composite process• Personal Agent: interface between user and EO• Dynamic input: Web form filled out by the user, user preferences• Conditional output: Complete itinerary with reservations booked and tickets purchased, or itinerary partially organized, or failure and error message• Future work: – DAML-S version advertising properties and capabilities through a profile description, and providing a grounding consistent with DAML-S 0.7 – Integration with a logic-based content language to be compatible with the requirements of semantic agent communication within Agentcities – Restaurant, cinema and taxi services
    18. 18. © 2002 Fujitsu Laboratories of AmericaChallenges (addressed by Agentcities)• Semantic interoperability, use of ontologies• Need for standard searching mechanisms and service classification• Tools, frameworks and methodologies, to help the software-agent developer• Dynamic, intelligent and autonomous composition of services
    19. 19. Availability Details (September © 2002 Fujitsu Laboratories of America2002) The Evening Organizer for San Francisco http://sf.us.agentcities.net/ Luigi Ceccaroni, Jonathan Dale lc@fla.fujitsu.comRestaurant services: http://sf.us.agentcities.net/aap/RestaurantServiceDescription.htmAAP on SourceForge.net: http://www.sourceforge.net/projects/networkagent

    ×