Human-Centered Systems for Business Services

Nabil R. Adam, Vijayalakshmi Atluri and Soon Ae Chun CIMIC, Rutgers University, Newark NJ 07102 {adam,atluri,soon}@cimic.rutgers.edu 1 Project Summary This NSF-funded project is a collaborative effort with the State of New Jersey to address two major problems in consuming and providing government services: the integration and customization of services. The government services needed by the citizens or businesses often require horizontal integration across autonomous government agencies, whose information systems are heterogeneous and difficult to interoperate. The service consumers, both citizens and businesses, face a challenging task of locating relevant services and information from various documents abundantly available on the Web. They are in need of a system to locate and integrate the services that are tailored to individual preferences and constraints. 2 Research Activities To guide entrepreneurs through the process of establishing their business, we have developed a decentralized workflow management system (WFMS) that is capable of (1) generating a customized workflow based on user parameters, requirements and constraints; (2) automatically executing the different processes involved at relevant agencies in the appropriate sequence by authorized individuals, adhering to the business policies of those agencies; (3) visually reporting the progression of the workflow to the entrepreneurs; and (4) integrating with existing state agency information systems. Our primary research activities to include: 1. Decentralized Workflow Execution Model: The decentralized workflow execution model eliminates the need for a centralized WFMS which has known problems such as limited scalability and performance bottlenecks. Our decentralized model relies on (1) on-the-fly construction of self-describing workflows and (2) their execution with the help of light weight WFMS stub located at each agency. A self-describing workflow is a partition of the workflow that is associated with the tasks’ definitions, the agencies responsible for their execution, and the control flow information. 2. Service Customization: Our service customization relies on a rule base reflecting the rules and regulations of the state agencies. In particular, our approach is based on the domain knowledge of services and regulatory rules, and is represented as service ontologies and rules topic ontologies. We have developed algorithms that evaluate compositional rules against a user profile, thereby automatically generating the inter-agency workflow customized to a specific user. 3. Dynamic Change Management: While a business registration process (in general, a workflow) is in progress, different changes may occur: (1) changes to the goals of the business process, that is, changes to the user requirements, (2) unanticipated exceptions that arise during execution, and (3) changes to the business rules and regulations. To handle these dynamic changes while delivering the eGovernment services, we have developed a decentralized workflow change management system that automatically identifies migration rules that specify operations and tasks needed to adjust the workflow to achieve run-time customization. 4. Ontology Construction: The rule base comprising of agency regulations is extracted directly from various state agency documents using Natural Language Processing technology. We have developed algorithms and a tool called OntoStruct, for automatic extraction of ontological information from text sources using domain ontologies and text processing technologies. This research has concentrated on finding relationships between terms, a set of services and regulatory rules. 3 Project Participants The collaborating state government agencies and representatives include Office of Information Technology (OIT): Wendy Rayner (former CIO), Judith Teller (new CIO), Adel Ebeid (CTO), Odysseus Marcopolus (Director, E-government), Sue Colbert (Manager, E-government), Lydia Quill, Lynn Raleigh, and Chris Foley;