Resource allocation is a complex problem as resources can have different competences, capabilities and rights. So, together with her colleagues, she posed the question:“What is the state-of-the-art of system-initiated resource allocation approaches for business processes?”

Some interesting findings:
- Papers may use models or execution data but rarely both
- Most papers decide based on previous performance and workload, but some include other interesting perspectives
- Most papers don’t deliver a prototypical implementation!

- Rule-based solutions are most used and have lowest execution cost,yet also propose the worst strategies
- Exact algorithms do the opposite: globally optimal solutions at the price of prohibitively long computations
- Approaches based on ML or (meta-)heuristics are in between

They also identified future work directions:
- Leveraging more execution data
- Exploring additional aspects of resource and task characteristics
- Increasing adaptability and flexibility of algorithms
- Benchmarking

Want to learn more? Then read the pre-print by Luise Pufahl, Sven Ihde, Fabian Stiehle, Mathias Weske & Ingo Weber here: https://arxiv.org/abs/2107.07264

P.S. Arxiv currently has an older version of the pre-print but it will be updated soon.

Remaining time prediction is an important area of PPM as it allows to check if a running instance will meet the time constraints and to act timely. Deep learning models are generally applied due to their higher accuracy, however, at the cost of explainability.

To this end, the authors propose feature-informed cascade prediction framework. It uses a novel explainable automated feature selection strategy and a deep learning model that allows to correlate the effects between each input feature and its prediction results.

First, the effect of each feature is evaluated in a systematic manner, resulting in a so-called incremental feature tree. The tree is pruned using a threshold, and a node with the smallest MAE has the best feature set on the path to it from the root.

Second, the feature-informed cascade prediction model is constructed. Each feature gets a separate layer, and layers are trained sequentially. The input sub-layer receives its feature and the output of the previous layer. The hidden sub-layer includes a neural network,e.g. LSTM.

In their evaluation on 8 real-life event logs, they show that FCPM performs better (lower MAE) than other approaches on most logs while having medium training time. Additionally, they showcase an end-to-end application scenario with a detailed case study.

The proposed framework achieves both higher accuracy and greater explainability. In future work, the authors plan to not only use more advanced deep learning models for improved accuracy but to also develop a process-oriented feature selection approach.

I especially like the explainability of the approach,which is key to eventual acceptance of the PPM (or any other) technology. And I agree with the authors that predictive models in context of business processes require additional process-specific features. So go read the paper!

BPMN is well known for its poor support for data flow.Despite attempts to improve it,some important aspects are still left out.The authors especially highlight issues when single process instance has to deal with multiple instances of data objects,especially at different states.

To this end, the authors propose semantics for multi-instance data objects and data object collections (DOC) and their translation to colored Petri nets. Authors define 4 operations for DOCs:
- State transitions
- Splitting
⁃ Merging
⁃ Creation

For instance, splitting a DOC into 2 collections of the same data objects in different states can be realized as shown below.
(See helpful explanations and the running example in the paper)

The authors see this work as starting point for future research on multi-instance behavior in process data flow. Further steps include more complex constructs,data locking mechanisms and derivation of BPMN modeling guidelines from the newly defined behavior. So go read the paper!

Personally,I have seen a lot of attempts to improve data flow support in BPMN,some of them even independently proposed by different authors.Still,they are not widely agreed upon neither adopted by major tools.Should we work towards BPMN 2.1 or 3.0?

Their paper presents different event log abstractions enabling LLM-aided business process analytics.
This approach yields inspirng results, especially for descriptive questions (see pic).
They also integrated it in PM4Py, so we all can benefit from their work already today!