Releases: AutomatedProcessImprovement/ongoing-process-state
2.0.2
2.0.1
Changelog
aca52fe - Merge Build&Test and Release in the same workflow (David Chapela de la Campa)
7cd777a - Update version (David Chapela de la Campa)
89e7646 - Fix 'release' workflow (David Chapela de la Campa)
26b4eee - Add workflow to automatically publish the package in PyPi (David Chapela de la Campa)
1b5fa39 - Refactor project/package name to "ongoing-process-state" (David Chapela de la Campa)
v2.0.0
Stable release with support to both sound Workflow graphs (BPMN) and sound Workflow nets (Petri net).
The performance of the Petri net implementation is better (up to 10 times faster when computing the reachability graph).
IEEE TSC v1.1.0a
Approach to, given a process model in BPMN formal, compute the state of ongoing cases in constant time. The approach consists of, in design time, given a maximum size n, create an index that associates each n-gram -- i.e., execution of n consecutive activities -- with the state(s) they lead to in the process model. Then, at runtime, the state of an ongoing process case can be computed in constant time by searching for the last n executed activities in the index. For example, for an ongoing case A-B-F-T-W-S-G-T-D, after building the 5-gram index, the state would be computed by searching in the index with the sequence [W, S, G, T, D].
This approach has been submitted as a publication to IEEE Transactions on Services Computing under the title "Efficient Online Computation of Business Process State From Trace Prefixes via N-Gram Indexing", by David Chapela-Campa and Marlon Dumas.
Instructions to reproduce the evaluation presented in the publication are available in the README file.
VLDB 2025 v1.1.0a
Approach to, given a process model in BPMN formal, compute the state of ongoing cases in constant time. The approach consists of, in design time, given a maximum size n, create an index that associates each n-gram -- i.e., execution of n consecutive activities -- with the state(s) they lead to in the process model. Then, at runtime, the state of an ongoing process case can be computed in constant time by searching for the last n executed activities in the index. For example, for an ongoing case A-B-F-T-W-S-G-T-D, after building the 5-gram index, the state would be computed by searching in the index with the sequence [W, S, G, T, D].
This approach has been submitted as a publication to VLDB 2025 under the title "Efficient Online Computation of Business Process State From Trace Prefixes via N-Gram Indexing", by David Chapela-Campa and Marlon Dumas.
Instructions to reproduce the evaluation presented in the publication are available in the README file.
v1.1.0
First stable release
IEEE TSC v2.0.0a
Approach to, given a process model in Petri net or BPMN format, compute the state of ongoing cases in constant time.
The approach consists of, in design time, given a maximum size n, create an index that associates each
n-gram -- i.e., execution of n consecutive activities -- with the state(s) they lead to in the process model.
Then, at runtime, the state of an ongoing process case can be computed in constant time by searching for the last n
executed activities in the index.
For example, for an ongoing case A-B-F-T-W-S-G-T-D, after building the 5-gram index, the state would be computed
by searching in the index with the sequence [W, S, G, T, D].
This approach has been submitted as a publication to IEEE Transactions on Services Computing under the title "Efficient
Online Computation of Business Process State From Trace Prefixes via N-Gram Indexing", by David Chapela-Campa and
Marlon Dumas.
Instructions to reproduce the evaluation presented in the publication are available in the README file.
ICPM 2024 v1.0.0
Approach to, given a process model in BPMN format, compute the state of ongoing cases in constant time. The approach has been submitted as a publication to ICPM 2024 under the title "Efficient State Computation for Log Animation and Short-Term Simulation Using N-Gram Indexing", by David Chapela-Campa and Marlon Dumas.
Instructions to reproduce the evaluation presented in the publication are available in the README file.