Business Process Modeling

This essay aims at discussing and elaborately expanding on the proceeds of the 11^th through the 13^th international conferences that were conducted on Business Process Management. The paper discussed has a direction relevance to business process management. The selection conference article for this essay is Case Analytics Workbench: Platform for Hybrid Model Creation and Evolution. Hybrid process models are perceived a luring approach that can be adopted on the modelling of processes that are knowledge intensive. A hybrid model is composed of both declarative and imperative processes which can deal with both the flexible and structured aspects of any business processes (Yu et al., 2015).

Nonetheless, it tends to be a challenge as well as consuming a lot of time to generate as well as purify a hybrid process owing to the sophisticated nature of the structure as well as the variability of the case. This paper is composed of an introduction to Case Analytics Workbench, a throughout system that is used in the acceleration of the process of hybrid process model as well an evolution through joining imperative and declarative process mining, human interaction and clustering in a cloud environment (Muthusamy et al., 2016). The validity of the effectiveness and applicability of the system is as well verified through conducting two case studies one from the healthcare sector and the other from insurance industry.

Business Process Modeling has a fundamental role to play when it comes to the domain of business process management that tends to be the activity that is representative of the processes in an enterprise to enable their execution, analysis as well as improvements. Besides the emerging knowledge intensive process, the imperative approaches comes across challenges (Ur Rehman, Chang, Batool & Wah, 2016). The carrying out and execution of the processes that are knowledge based mainly rely  on the knowledge staff working on different linked to decision making that is knowledge intensive responsibilities that are centric to information and need elaborate flexibility. Approaches of declarative approaches among them DCR Graphs, Declare and SCIIFF have been recently suggested to be in support of the required flexibility. Just the mainly features are discussed in the model way the declarative models while the constraints between the various activities are defined explicitly to inhibit banned behaviors (Yu et al., 2015).

A research among the various scholars illustrate that the declarative language of modeling is not among the process modelling languages which can be adopted in the whole process processes in real practice. It as well demonstrated that in a number of the mechanisms, at least a fraction of the processes are better substituted through the use of imperative approach. As a general perception, the hybrid modelling approach, in combination with the declarative and imperative modelling approaches, tends to be a more luring way of modelling the whole business processes (Rosemann & vom Brocke, 2015).

From the human effort perspective, the creation and refining of a hybrid process does not qualify to be a trivial work specifically for knowledge intensive mechanisms. In the first place, the modeller should be equipped with the general perspective regarding process model. It would turn out to be more challenging besides the ever increasing complexity of information or structure of the process. The second aspect involves having an understanding or otherwise learning curve for the modeller so as to get hold of new hybrid modelling guidelines and languages (Yu et al., 2015). The third aspect involve ensuring that the model hybrid is kept up to date and this involves the modeller being in a position to run the methods of process mining and elaborating the results extensively for the purposes of refining the model. The last aspect would involve the modeller calling on the attention of other analytics methods in the provision of customized modeling.

Research Method Used

Case Analytics Workbench is introduced as an end to end system that may be adopted in easing the workload of process modellers as well as accelerates the creation and evolution of hybrid process model. The system is such that it:

• Leverages event log clustering in offering support to the reorganized process modelling
• Brings together the results of imperative and declarative process mining in the extraction of data driven evidence as well synthesizing them automatically (Yu et al., 2015)

This research adopted a report on the case of two real world use, that are derived from the health care industry and insurance in the demonstration and evaluation of the ability of the workbench hat is proposed. The first case described how the various compositions in the workbench are orchestrated in a way that can aid the modeller in the creation of an underwriting process model in a manner that is most convenient. The second case study on the other hand, explores on the refines of sophisticated hybrid model (Yu et al., 2015).

The case study has the server side of the workbench being adopted on the IBM Bluemix, that tends to be a platform that is cloud-based that is used for building, operating as well as management of various apps. The components  of database of the server-side  of the workbench were used a Bluemix Services through the use of SQL Database Service in  Bluemix beside Cloudant as well as other parts were used as Bluemix APPs that offered RESTful APIs.

The case study structured in such a way that it served as an optimization and improvement of the underwriting process which can be traced back to the real specifications of an insurance company (Yu et al., 2015). This involved the creation of a raw case model as illustrated in the figure below in which the primary skeleton as composed of activity as well as structural elements including Task, Case, Step as well as Stage elements

Preparation and Clustering of Data: This involved the collection of 4300 historical execution logs from the insurance company of the undergoing process that lasted from August through December 2013. The collected data was thereafter converted to firma event logs through the use of data preparation aspect and thereafter the triggering f event log clustering so as to come up with definite from the initial model (Paige, Matragkas & Rose, 2016). This led to the generation of two clusters, one containing 2038 case instances and the other 2267 instances. The visualization of the results of the declarative process mining is as shown in the figure below

Figure 1: Visualization of the results of the declarative process mining: Extracted from (Yu et al., 2015)

Process Mining: Figure 1 is illustrative of the findings of the declarative processing mining technique which as well adopted the raw case model as the input and had Cycle time less than two day as the set goal (Balko & Vasudevan, 2014). All the findings were visualized in a Dendrogram to enhance elaborate understanding since the Dendrogram illustrated clearly the structures of the types of variation. Each of the findings was llustrated as a leaf of the Dendrogram node with its key features among them support degrees as well as a correlation with the goal attainment. It was noted that some important and reasonable elements, not identified in the raw model among them extra tasks besides extra constraints were successfully realized by the declarative process-mining motor.

Case study 1: Derivation and Creation of Underwriting Process

Evidence and Evidence Synthesis: The pieces of evidence illustrated in figure 2 were changed from findings of both imperative and declarative engines of process mining. Strong evidence having a support of greater than 0.3 as well as a goal correlation of more than 0.5 was chosen through the filter before the automatic evidence synthesis engine was run. Figure 3 shows the UI synthesis (Vom Brocke, Petry & Gonser, 2016).

Figure 2 (a) The synthesis overlay on the raw case model; (b) The final model after synthesis: Extracted from (Yu et al., 2015)

This involved carrying out a case study on care pathway refinement. A care pathway defines a standardized channel which is composed of numerous care stages that are in correspondence with various conditions of disease progress, in which each of the stages is composed of different clinical tasks as well as their constraints (Yu et al., 2015). Owing to its strong demands of hoc variation and flexibility, it is supposed that the hybrid approach would be ideal in modelling the complex care pathways. This study involved building an initial care pathway manually as a case model through the use of the workbench. This was a derivative of the clinical guidelines for congestive heart failure management. The model was then refined for a given patient cohort pegged on the mined pieces of evidence from the actual electronic medical records.

The engines for process mining offering important proofs that were used for the purposes of making improvements on the care pathway. An example of such was witnessed in the initial care pathway that was extracted from the guideline in which some of the constraints outlined that the baseline tests had to be carried out before the initiation of the process of treatments. Nonetheless, as illustrated in the figure below, there was detection by the declarative process mining engine that the treatments were often initiated without carrying out baseline tests in the actual electronic medical records data. A concrete explanation is offered by clinicians regarding such a violation.

Figure 3: The care pathway model, selected evidences and the synthesis findings: Extracted from (Yu et al., 2015)

In actual practice, clinicians offered treatments following the ancient medical histories which might not have been captured in the electronic medical records hence such constraints are not needed and thus can be assumed in some of the cases. Still, the findings of the care pathways mined through the process mining strategy were applied. Through the actions of treatment explained in the initial case model, fragments were chosen manually from the care pathway model and the likely consequences added into the refined model (Thramboulidis, Vachtsevanou & Kontou, 2018).

Conclusion

A solution to the acceleration of the creation and evolution of the hybrid process model is presented in this research paper through a combination of both the imperative and declarative process mining using event log clustering. The research built up a throughout system called Case Analytics Workbench in the cloud surrounding to enable subscribers do a check analysis of the findings as well as elate with the various case models, besides the actual work case studies drawn from the health care industry and the insurance.

From the perception of design, cloud based architecture attains extendibility and adaptively for various use case design while for the point of applicability, the real world case studies as well as their findings made firm effectiveness to experts among them business process administrators, BPM product developers as well as managers, BPM administrators besides clinical physicians. A majority of them returned positive results regarding a combination of the wide range of process mining techniques besides the inclusion of clustering method. They as well unanimously came to a consensus  that the processing of generating and enhancing a business process model ought to be inclusive of the user interaction as well as the mechanisms of user relation among them visualization of the results of analytics as well as the design  of the model of interaction was acknowledged by the experts. There are a number of further considerations that should be made with regard to hybrid process modelling. One of such is owing to the inadequacy of CMMN and BPMN, there was a clear distinction of the declarative part from the imperative part in the definition of the case model which turned out to successively work for the case studies. In some other scenarios, there may be an insignificant distinction between the model elements in which there were two modelling techniques should be adopted and thus calling for a more flexible model of hybrid process to be designed.

References

Yu, Y., Li, X., Liu, H., Mei, J., Mukhi, N., Ishakian, V., Xie, G., Lakshmanan, G.T. and Marin, M., 2015, August. Case analytics workbench: Platform for hybrid process model creation and evolution. In International Conference on Business Process Management (pp. 226-241). Springer, Cham

Muthusamy, V., Slominski, A., Ishakian, V., Khalaf, R., Reason, J. and Rozsnyai, S., 2016, June. Lessons learned using a process mining approach to analyze events from distributed applications. In Proceedings of the 10th ACM International Conference on Distributed and Event-based Systems (pp. 199-204). ACM

ur Rehman, M.H., Chang, V., Batool, A. and Wah, T.Y., 2016. Big data reduction framework for value creation in sustainable enterprises. International Journal of Information Management, 36(6), pp.917-928

Paige, R.F., Matragkas, N. and Rose, L.M., 2016. Evolving models in model-driven engineering: State-of-the-art and future challenges. Journal of Systems and Software, 111, pp.272-280

Oestreich, T.W., 2016. Magic quadrant for business intelligence and analytics platforms. Analyst (s), 501, p.G00275847

Thramboulidis, K., Vachtsevanou, D.C. and Kontou, I., 2018. Cyber-Physical Microservices and IoT-based Framework: The case of Evolvable Assembly Systems. arXiv preprint arXiv:1807.07363

Vom Brocke, J., Petry, M. and Gonser, T., 2016. Business process management. In A Handbook of Business Transformation Management Methodology (pp. 137-172). Routledge

Balko, S. and Vasudevan, K., SAP SE, 2014. Business process management. U.S. Patent 8,849,747

Rosemann, M. and vom Brocke, J., 2015. The six core elements of business process management. In Handbook on business process management 1 (pp. 105-122). Springer, Berlin, Heidelberg