FLEXIBLE PROCESS MODELING THROUGH DECISION TABLES Nicola Boffoli, Daniela Castelluccia, Fabrizio Maria Maggi, Roberto Rutilo Department of Informatics – University of Bari - Via Orabona, 4, 70126 Bari – Italy email: {boffoli, castelluccia, maggi, rutilo}@di.uniba.it ABSTRACT As today enterprises have to face a competitive, complex and ever-changing business context, so business processes have to respond to the context changes in a fast and economic manner. This purpose implies that processes should be continuously maintained through a flexible modeling. This paper addresses this problem and provides an approach able to govern the high variability of the environment where the process performs. This approach is based on the well-known formalism of decision tables to link context parameters to process activities, so that it allows a fast process re-modeling when context parameters change. We have tested this approach in real contexts: results are encouraging and drive further investigations in such a way. KEY WORDS process modeling, context flexibility, decision tables 1. Introduction Nowadays enterprises perform in an extremely competitive business context, therefore business processes, although complex, must be highly flexible to react to the new business demands. Any change of the business environment, represented by workers, tools, cultural factors, industrial standards, quality programs and budget, directly impacts on the adequacy of the business processes and, in general, on the responsiveness of the enterprise [1, 2, 3]. Therefore the dynamism of the environment, where a business process performs, is the main factor that mostly increases the need to modify and adapt the processes. That’s why Business Process Management (BPM) is spreading: in fact it provides methodologies for business process modeling, monitoring and continuous improvement in order to govern the process complexity and the environment dynamism. According to BPM methodology, business processes are represented through highly intuitive models where a set of activities are executed by a set of actors (human or automatic), interacting with different information sources and enabling further established activities. Although BPM systems aim to increase process flexibility through the composition of sub-processes, the most BPM systems lacks a real support for process modeling related to environment dynamism. For instance, process patterns in [4, 5, 6] define how to arrange activities inside the process (task sequencing, split parallelism, join synchronization, iteration) but they don’t identify a connection between the process and the changing context. In [7, 8] process patterns are specialized according to the context but this specialization task is at design level, not at analysis and modeling, so that the process reengineering is complex and not suitable to support context changes. In order to increase flexibility, it’s necessary to provide a clear representation of the connections between the context and the process solutions. Moreover it’s necessary to support the decision making during the composition of a process solution according to an established business goal into a determined operative context [9]. In recent years, other authors have addressed this problem providing different methodologies for process flexibility related to the environment changes. In [10] a flexible modeling method, based on the capability of extensible organization description, is proposed. In [11, 12] the authors represent the connections between processes and contexts using ontology-based semantics. In [13] connections between contexts and process solutions are analyzed firstly, then a business modeling methodology based on business and analysis patterns is proposed. Finally, in [14] the flexibility is supported through the Role of Cased-based Reasoning to solve a new problem remembering a previous similar situation and reusing its information and knowledge. Instead the authors of this paper provide a methodology based on the use of decision tables in order to identify a process solution taking into account the specific context where that solution will perform. Decision tables assure a compact overview of a large number of information, a modular knowledge organization, an effective evaluation of consistency, completeness and redundancy. These peculiarities guarantee a representation of the relationships between contexts and process solutions in a complete manner, without inconsistencies and fast reusable. That’s why decision tables allow to identify all the possible contexts for each process solution and to reject the inconsistent ones. Moreover decision tables are easily maintainable. This fact increases flexibility of dynamic reengineering of the relationships between contexts and process solutions: when the environment changes, it’s possible to identify the new specialized solution just changing some elements of the decision tables and easily to change the process replacing the existing solution with a specialized one according to the new context. 598-182 280