1	Enhancing the Intelligence of Mobile Middleware Environment Weishan Zhang School of Software Engineering, Tongji University, Shanghai, China Thomas Kunz Systems and Computer Engineering Carleton University, Ottawa, Canada
2	Outline Motivation Context ontologies in the user-centered mobile computing Combining the power of BDI agents and OWL ontology reasoning mechanism Outlook
3	Motivation User-centered applications in mobile environments Location-based services Personalized home entertainment Social coordinator (schedule appointments, contact friends) etc. à access to and use of context information Mobile devices: resource-constrained (computational power, memory, I/O, …) many different platforms wireless access technologies (highly variable bandwidth, intermittent connectivity, …..)
4	Motivation (cont.) Rich body of work on context and ontologies in semantic web Deal with platform/network heterogeneity: middleware Review of existing middleware: No mechanisms to handle ontology evolution resulting from the timely adaptation of an ontology to changes Weak intelligence: mobile agents are only used as an agent for code mobility; no intelligence is built into the agents.
5	Context ontologies in the user-centered mobile computing
6	PERSON ontology
7	Context ontologies in the user-centered mobile computing
8	Basic Idea Combine BDI agents with OWL ontology reasoning BDI agents: weighing conflicting considerations provided by what the agent desires/values/cares about and what the agent believes (practical reasoning) OWL ontology reasoning: generate new beliefs from ontology information (theoretical reasoning) In addition: apply frame concept/XVCL to ontologies to handle ontology evolution
9	Combining the power of BDI agents and OWL ontology reasoning mechanism
10	Combining the power of BDI agents and OWL ontology reasoning mechanism According to Jadex architecture, this is theoretically feasible using the dynamic model of Jadex
11	Frame-based Java Ontology Processing Bridging the world of OWL ontology and Java software agents. It uses the advanced template mechanism of Frame technology to map the concepts and properties in OWL to the related Java class and attributes. Management of ontologies and handling of ontology evolution based on meta-ontologies developed with XVCL. Management of the agent definition updates, including the agent belief sets, plans and goals.
12	A case study of context-aware person agents negotiating sport appointments
13	BDI Agents Implemented with Jadex beliefs, goals and plans are first class objects in a JadeX agent implementation components are an XML agent definition file-ADF (where Beliefs, Goals and Plans are modeled with XML tags) Java code for the implementation of agent actions for the Plan. Achieving sports negotiation goal straightforward: Initiator first searches the online person agents, verify whether they are in the list of friends sharing the same hobby sport. Then, the proposal will be sent to all online people sharing the hobby, etc. “Missing”: how to determine who shares the same hobby sport, subject to constraints such as “live in the same city” or others
14	Query for answer hobby sharing
15	Example of enhancing beliefs during ontology evolution
16	Example of handling of ontology evolution (adding concepts)
17	Conclusions and future work Proposed a combination of BDI agents and Ontology Reasoning to develop user-centric mobile applications Built prototype of proposed framework Since we use frame concept, can handle ontology evolution seamlessly Future work: Implement BDI agent in J2ME version of Jadex, and agents based on Jade-LEAP for small devices which have no easy access to reasoning capabilities. We will continue the implementation of the FOJP components to finish the mapping from the Ontology model to Java classes based on the Frame based template mechanism, which is currently done manually.
18	"Thank you.." Thank you..