Integrated business process adaptation towards friction-free business-to-business collaboration

摘要

One key issue in process-aware E-commerce collaboration is the orchestration of business processes of multiple business partners throughout a supply chain network in an automated and seamless way. Since each partner has its own internal processes with different control flow structures and message interfaces, the real challenge lies in verifying the correctness of process collaboration, and reconciling conflicts in an automated manner to make collaboration successful. The purpose of business process adaptation is to mediate the communication between independent processes to overcome their mismatches and incompatibilities. This dissertation proposes a new framework and efficient algorithms for creating integrated adapters for business process composition. For the control flow adaptation, we propose a structural analysis of patterns which we show is more efficient than existing approaches. Furthermore, we propose algorithms based on message pair analysis and integer programming to create an optimal adapter in both synchronous and asynchronous communication. These algorithms are designed to handle more general cases involving M messages and N processes (M, N > 2). For message adaptation, we identify a set of extendible message adaptation patterns to solve typical message mismatches. Finally we show algorithms to integrate individual message adaptation patterns with control flow adapters to create a complete adapter for multiple processes. All algorithms were implemented in a Java-based prototype system, and results of validation test and performance experiments are reported. We compare and discuss the insights gained about adapter creation in different scenarios.