Modeling, analysis, and composition of business processes

  • Authors:
  • Jianwen Su;Oscar H. Ibarra;Cagdas Evren Gerede

  • Affiliations:
  • University of California, Santa Barbara;University of California, Santa Barbara;University of California, Santa Barbara

  • Venue:
  • Modeling, analysis, and composition of business processes
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Service Oriented Architecture (SOA) is centered on the idea of using services as the building blocks for application design, development, and management. This paradigm has painted a bright picture for future software development. Much of the expected benefits come from a systematic approach of sharing program executions (i.e., services), as well as data. One of the domains in which SOA can have a big impact is Business Process Management. A business process is a set of activities performed to accomplish business objectives. The goal of Business Process Management is to provide tools to define, analyze, execute, and monitor business processes in order to achieve continuous improvement of businesses. Today the management of business processes is becoming more difficult. First, increasing competition forces businesses to respond to changes more rapidly and utilize available resources maximally when they need to provide new services. As a result, there is a shift happening toward building new services on top of existing services. Second, changes in the business environment happen more frequently and current business process modeling approaches have problems dealing with change. In this dissertation, we first address the problem of composing business services from existing services automatically. Based on the Roman service model, we develop algorithms for the automated service composition problem. Our analysis considers various cases including services with finite states, services with bounded states, and services with qualities. Second, we investigate an alternative approach to business process modeling, namely artifact-centric business process modeling. This approach promises to address the problems of current process modeling approaches by emphasizing data design as well as control flow design. Through our collaborations with IBM Research, we have been able to refine and formalize this approach. Based on our formalization, we develop a logic based language to specify life-cycle properties of artifacts and show various decidability results on the verification of these properties. Our results contribute to the foundational knowledge required for the development of new tools addressing today's business process management challenges.