Heureka!

I am developing two different heuristics to translate and fix process models from a poor quality to a good quality. It should be possible to provide a generable way to improve process models by using heuristics, hence heureka!

The process models I retrieved from the hospital are all modeled in Visio, which makes these processes static, not-executable, and hard to analyze. To be able to analyze these models, they should have some equivalence in the used labels and in the used perspective. Therefore I came up with a ‘translate heuristic’ that translates process models from any sort to BPMN based process models. The result of this heuristic are BPMN based process models. However, some process models are modeled more extensive than others. In other words, there are (huge) differences between the process models from one department compared to the models from another department. To level out these differences, I created another heuristic; the ‘fix heuristic’.

The ‘translate heuristic’ is used to translate flowchart models to a BPMN based models, and I think this heuristic is generalizable to another type of models as a starting point. Imagine you have Event-driven Process Chains (EPC) based model as a starting point. By following the same heuristic, a translation from EPC to BPMN should be possible.

The ‘fix heuristic’ takes the BPMN based process model as a starting point and fixes all/most of the errors, at least that is why I have been developing this. To solve some errors you should involve someone that has domain knowledge and/or expertise of that particular process, while other errors can be solved without domain knowledge.

I have not published these heuristics (yet) because they are work in progress, but you can provide me with some of your precious feedback and I can take that into account when developing/improving these heuristics.