Rule-based business process modeling and execution

A process model is called rule-based if the semantics of its case data and activity flows are expressed by means of rules. Rules have been recognized before as powerful representation forms that can potentially define the semantics of data and process resources. To date, however, there is no consensus on how to link the enforcement of rules, the manipulation of data and the execution of processes. Moreover, it is witnessed that complex data and process resource descriptions in the form of a number of constraints, deduction and reaction rules lack expressivity and comprehensibility. In this paper, we set out for using process and rule set metamodels to concisely represent the semantics of case data and activity flows in business process models. In addition we show how to generate a syntactically verified and semantically validated corpus of definite Horn clauses that can be used in the execution of the modeled process. I. I NTRODUCTION Software engineering aspires to avoid the duplication of resources. This is a fundamental principle which is based on the experience that systems with duplicated resources sooner or later run across a myriad of difficulties. As with data in the past and process descriptions at present, logic is gradually becoming the next resource to be managed outside individual applications. Through the separation of so called business rules from applications, it is hoped that changes in business logic will no longer result in an avalanche of required application updates, and will thus reduce the IT bottleneck when bringing about business policy changes. Separating logic from applica- tions is the goal of rule-based software engineering , which has the potential of significantly improving the theory and practice of system design. This separation, however, raises the problem of how to link the enforcement of business rules, the manipulation of data and the execution of processes. Several approaches to this problem are described in the literature, such as Dietrich's rule-based agents (1) and D'Hondt's approach which considers business rule enforcement as an aspect-oriented programming cross- cutting concern (2). In this paper, business rule enforcement is situated at the level of business processes rather than at the level of individual applications. In particular, we propose a process and rule set ontology for rule-based business process modeling and an architecture for rule-based business process execution. A. Existing process languages