Summary: An Algorithm for Business Process Compliance

G. Governatori and A. Rotolo. An algorithm for business process compliance. In E. Francesconi, G. Sartor, and D. Tiscornia, editors, JURIX, volume 189 of Frontiers in Artificial Intelligence and Applications, pages 186–191. IOS Press, 2008.

Business process compliance means the adherence or consistence of a set of specifications modelling a business process and a set of specifications modelling the norms for a particular business. Business process specifications describe how a process is executed, while norms state what can be done and what cannot be done by a process. Compliance checking is to figure out (a) which obligations will definitely appear when executing the process, and (b) which of those obligations may not be fulfilled. This paper introduces an algorithm that provides a compliance checking.

The paper first chooses FCL as the formal rules language, FCL (Formal Contract Language) is a logical language that represents rules as expressions. FCL uses the following operators ¬(negation), O (obligation), P (permission), and ⊗ (violation/reparation), ∼p denote the complement of p. (≺) determines the priorities between two rules to avoid conflicts. FCL expressions such as:

r1 : account(y) ⇒ OpositiveBalance(y) ⊗ OapproveManager(y)

Which means for account y it is an obligation to have a positive balance, if this obligation was violated then it is an obligation to get manager approval. After writing all rules in FCL, one can enrich the process model by these rules expressions.

Then the paper explains process models showing how to annotate them with these rules expressions. Finally, it presents the compliance-checking algorithm. To do compliance checking there are three steps to go through. Step one ‘building procedure’, consists of the following: 1- identifying sets of effects for each task from the process model. 2- determine the normative position (obligation, permission, prohibition) for each set of effects of these tasks. 3- for each task check if obligation is fulfilled or a violation has occurred. Step two ‘from tasks to obligations’, which determine the obligations derived by the tasks. Step three ‘checking compliance’, is applying the checkCompliance algorithm.

Governatori and Rotolo in this paper proposed a compliance algorithm, using FCL as the rule modeling language. FCL has the ability to represent obligation, permission, and negation. It also has the ability to show what to do incase of violation. And also uses a priority operator to avoid conflicts. FCL do not differentiate between different types of obligations (e.g. if an obligation is continues or occur once). Also it does not have the ability to represent time constrains regulations, or if an obligation need to happen after another obligation already happened and never before.