Abstracting Common Business Rules To Petri Nets

Business information systems are mostly very complex and difficult to modify. As a consequence, if we would like to ensure that certain business rules are enforced in a business, it is often easier to design a separate information system, called a monitor, that collects the business events from the business information system in a log and verifies whether the rules are satisfied. If not, the monitor can report the problem or it can intervene in the business information system. This requires that the business rules are specified in a precise language that allows to verify them over log files. In addition it should use a vocabulary that is familiar to business experts, and use terms such as cases, resources, agents and tasks. We introduce such a language and show that it can readily express many common types of business rules. We do not need the whole log but only an abstraction to those (aspects of) events that are relevant for the business rules. In addition we show that in many cases these rules can be translated to Petri nets which gives us not only an efficient way to detect violations of business rules, but also a way to analyze them. Such as Petri net can be seen as an abstraction of the business rules. The latter enables us to detect events that either may or must lead to violations later on. This way the monitor can report the risk of transgressions or even prevent them by intervening in the execution of the business information system.