Business Process Instances Research Articles

Predictive process monitoring: concepts, challenges, and future research directions

Predictive Process Monitoring (PPM) extends classical process mining techniques by providing predictive models that can be applied at runtime during the execution of a business process, for example, to predict the sequence of the next event(s) in a case, its outcomes, or performance-related aspects such as the remaining processing time. These predictive models go beyond process mining’s inherent descriptive nature that is offered by typical process discovery, conformance checking, and model enhancement techniques. The growing interest in PPM is driven by its additional value proposition, i.e., delivering real-time information regarding the future execution of business process instances, thus allowing for better informed operational decision making and performance analysis. The rapid growth of PPM during the last ten years has left the field lacking a cohesive taxonomy and an explicit recognition of the prevailing challenges. This paper aims at closing this gap, by comprehensively defining PPM using a unified approach to the key terms and by discussing challenges and opportunities in the field. Specifically, we propose three overarching research challenges and nine research directions, which have been validated through a survey with PPM researchers.

When to Invoke a Prediction Service for Business Process Monitoring?

Predictive monitoring of business processes aims at forecasting the future information of a business process and has gained increasing attention in recent years. With the development of cloud computing, prediction models, including remaining time prediction models, can be provided as cloud services. Reducing the number of invocation times of a remaining time prediction service is necessary since a large quantity of business process instances may be initiated and monitored every day. However, most of the current research focuses on designing new algorithms to improve prediction accuracy but there is no approach available to decide when to invoke a prediction service for each business process instance. In this paper, we propose a deep reinforcement learning based strategy that can learn the policies of selecting prediction points for the remaining time prediction. Specifically, the learned policies can dynamically decide the next prediction point at which the remaining time prediction service will be invoked for a business process instance. We performed extensive experiments on five real-world datasets. The experiment results show this strategy can reduce the number of prediction points significantly and still maintain the high prediction accuracy.

Predictive Business Process Monitoring Approach Based on Hierarchical Transformer

Predictive business process monitoring is predicting the next stage of the business process based on the sequence of events that have occurred in the business process instance, which is positive for promoting the rational allocation of resources and the improvement of execution efficiency. There are drawbacks in modeling business process instances, such as conceptual drift phenomenon and long event sequences. Therefore, we propose a hierarchical Transformer-based business process prediction model to improve the performance of the Transformer-based predictive business process monitoring model. We encode the event features using two different encoding methods to obtain the relationship between activities and attributes. A drift detection algorithm is proposed to segment the business process and calculate the correlation between activities and segments by using cross-attention. Furthermore, learnable position encoding is designed to capture the relative position information of subsequences. Finally, the information of different granularity, such as event attributes, event subsequences, and complete instances, is fused by different weights. Experiments were run on seven real event logs for the next activity prediction and remaining time prediction, and the next activity prediction accuracy improved by 6.32% on average, and the mean absolute error of remaining time prediction reduces by 21% on average.

Software business process adaptive approach supporting organization architecture evolution

AbstractSoftware maintenance and evolution play an important role in the software engineering field, especially when current software becomes more and more complex and powerful. As an entity to implement business processes and gain revenue, valuable software is composed of business logic and corresponding organization role interaction interfaces. With the enterprise development, the organization architecture also evolves, like expanding, cross department cooperation, and so on. However, existing software process adaptive approaches mainly focus on handling the change of the business (program) logic instead of organization structure. Therefore, we propose an adaptive software business process approach that supports organization architecture evolution and automatically migrates the run‐time process instances to the latest version. First, a business process adaptation model is designed, which includes the organization layer, business process layer and event layer that connects the two. Based on the model, the organization changing impact and business process model modification are formalized. Besides, the business process adaptation approach is designed. According to the dependence between the organization architecture and the business process activities, the affected domain detection algorithms for three basic business process structures and the business process instance migration algorithm are developed. Finally, the feasibility and stability of the proposed system are comprehensively evaluated with the synthetic data sets.

XAI in the Context of Predictive Process Monitoring: An Empirical Analysis Framework

Predictive Process Monitoring (PPM) has been integrated into process mining use cases as a value-adding task. PPM provides useful predictions on the future of the running business processes with respect to different perspectives, such as the upcoming activities to be executed next, the final execution outcome, and performance indicators. In the context of PPM, Machine Learning (ML) techniques are widely employed. In order to gain trust of stakeholders regarding the reliability of PPM predictions, eXplainable Artificial Intelligence (XAI) methods have been increasingly used to compensate for the lack of transparency of most of predictive models. Multiple XAI methods exist providing explanations for almost all types of ML models. However, for the same data, as well as, under the same preprocessing settings or same ML models, generated explanations often vary significantly. Corresponding variations might jeopardize the consistency and robustness of the explanations and, subsequently, the utility of the corresponding model and pipeline settings. This paper introduces a framework that enables the analysis of the impact PPM-related settings and ML-model-related choices may have on the characteristics and expressiveness of the generated explanations. Our framework provides a means to examine explanations generated either for the whole reasoning process of an ML model, or for the predictions made on the future of a certain business process instance. Using well-defined experiments with different settings, we uncover how choices made through a PPM workflow affect and can be reflected through explanations. This framework further provides the means to compare how different characteristics of explainability methods can shape the resulting explanations and reflect on the underlying model reasoning process.

Predicting remaining execution time of business process instances via auto-encoded transition system

As an important task in business process management, remaining time prediction for business process instances has attracted extensive attentions. However, most of the traditional remaining time prediction approaches only take into account formal process models and cannot handle large-scale event logs in an effective manner. Although machine learning and deep learning have been recently applied to the remaining time prediction task, these approaches cannot incorporate domain knowledge naturally. To overcome these weaknesses of existing studies, we propose a remaining execution time prediction approach based on a novel auto-encoded transition system, which can enhance the complementarity of process modeling and deep learning techniques. Through auto-encoding the event-level and state-level features, the proposed approach can represent process instances in a comprehensive and compact form. Furthermore, a transfer learning strategy is proposed to train the remaining time prediction model so as to avoid overfitting and improve the accuracy of prediction. We conduct extensive experiments on four real-world datasets to verify the effectiveness of the proposed approach. The results show its superiority over several state-of-the-art approaches.

Business Process Event Prediction Through Scalable Online Learning

Predictive process monitoring techniques aim to forecast outcomes of running business process instances. These techniques are based on using predictive models built from past observed behavior, i.e., in an offline setting. However, business process behavior usually changes over time and predictive models are therefore at risk of becoming obsolete. Because of this, the definition of systems that build predictive models through an online setting has recently gained attention. Nevertheless, the scalability of this kind of setting within a context where the amount of data available is experiencing rapid growth is an outstanding issue. To solve this problem, this paper aims to define a framework for event sequence prediction capable of taking advantage of modern distributed processing platforms. An implementation over this framework based on Apache Flink is presented and it is tested upon two different case studies to prove its validity and its capacity to scale.

Changes in artefact‐centric business process instances and their correctness prediction

Change of artefact-centric business process instances is important for an enterprise to keep competitive. However, to preserve the correctness of changes of an artefact-centric business process instance is still a big challenge. The hardness mainly stems from the fact that whether a business process instance can reach a final state has been proved to be undecidable. As such, finding a reliable verification algorithm for preserving correctness becomes impossible. In this study, the authors propose a random-forest-based approach to predict the correctness of changes in a business process instance. The availability of the proposed method is validated by comparing it to the traditional formal verification method. They also propose two optimisations of the proposed method and validate their effectiveness through extensive experimental analysis.

Cause vs. effect in context-sensitive prediction of business process instances

Predicting undesirable events during the execution of a business process instance provides the process participants with an opportunity to intervene and keep the process aligned with its goals. Few approaches for tackling this challenge consider a multi-perspective view, where the flow perspective of the process is combined with its surrounding context. Given the many sources of data in today’s world, context can vary widely and have various meanings. This paper addresses the issue of context being cause or effect of the next event and its impact on next event prediction. We leverage previous work on probabilistic models to develop a Dynamic Bayesian Network technique. Probabilistic models are considered comprehensible and they allow the end-user and his or her understanding of the domain to be involved in the prediction. Our technique models context attributes that have either a cause or effect relationship towards the event. We evaluate our technique with two real-life data sets and benchmark it with other techniques from the field of predictive process monitoring. The results show that our solution achieves superior prediction results if context information is correctly introduced into the model.

Prediction of Incompliance With Business Goals With Business-Related Data and Context Data

Proactive adaptation of business processes can prevent and mitigate future problems during the execution of business processes. For the proactive adaptation of business processes, adaptation decisions should be made based on whether a business process instance is heading to a business goal. Recently, much work struggles to improve predictive business process monitoring efficiency, i.e., accuracy and prediction earliness. As such, process designers can find more necessary adaptations and have more remaining time for the adaptations. Almost all of these previous efforts concentrate on business-related data generated by business processes. However, besides business-related data, context data has a huge impact on how a business process instance is unfolding to its completion. Context data should be noticed in the area of predictive business process monitoring. To remedy this shortage, in this paper, we propose a predictive business process monitoring framework based on a joint of business-related data and context data to predict incompliance with business goals. We evaluated the framework concerning the measures of accuracy, prediction earliness, and cost time based on a real-life online banana purchase process.

The problem of effective deployment architecture in SOA

Service Oriented Architecture is popular in many organizations. In particular, it has already deeply rooted in large corporations that need to automate entire business processes and implement them in many systems. It has a unique feature that allows unambiguously indicate service that is to realise business process step. That indication is possible to show directly in BPMN diagram. Thus, it is possible to trace which server has used resources to implement the service and how much of those resources were needed. Therefore, it is possible to build an optimization task that, with limited and unreliable resources, will determine such allocation of components to servers and such an algorithm for assigning tasks to them, so that the processes will work as well as possible. The article presents a model of such an optimization task. This model consists of four layers. The Organization Layer describes the system environment – the types and frequency of initiating business process instances. The Integration Layer describes the business processes and indicates the services that should be performed at every step. The Component Layer describes component characteristics and what services they provide. In Server Layer both: server characteristics and runtime environments necessary for the component to run are described. Finally, the optimization task and evaluation criteria are formulated.

Real-time Prediction and Synchronization of Business Process Instances using Data and Control Perspective

Nowadays, in a competitive and dynamic environment of businesses, organizations need to moni-tor, analyze and improve business processes with the use of Business Process Management Systems(BPMSs). Management, prediction and time control of events in BPMS is one of the major chal-lenges of this area of research that has attracted lots of researchers. In this paper, we present a4-phase pipeline approach to the problem of synchronizing each pair of dependent process instancesto arrive at the corresponding pair of tasks simultaneous or near-simultaneous. In the rst phase,the process model is mined from the event log and enriched by the probabilistic distributions oftime information. In the second phase, the hidden processing dependency between the each pair ofdependent process instances is formally de ned and is mined from the event log. In the third phase,a process state prediction algorithm is presented to predict the future route of process instance andthen predict the remaining time of the process instance to a given point in a predicted route of thebusiness process. In the fourth phase, an iterative synchronization algorithm is presented based onthe presented process state prediction algorithm to make each pair of dependent process instancesarrive at the corresponding pair of tasks simultaneous or near-simultaneous. Experimental resultson a real-life event log of BPI challenge 2012 show that the proposed method leads to 39% reductionin cycle time for dependent process instances.

The Adaptation of a Model of an Artifact-Centric Business Process Instance and Its Validation

The adaptability of an in-progress business process is an essential requirement for any business process management system in dynamic business process environments. Over the last two decades, the artifact-centric approach for business process management has been evidenced to have higher level of flexibility. However, the adaptation of a model of an artifact-centric business process instance is still inevitable and pervasive due to the complex and ever-changing business environments. Almost all works of artifact-centric business process neglect this issue. To fill this gap, we propose a special business rule called adaptation rule to address the dynamic adaptation problem and describe the adaptation by a global adaptation model. Moreover, we provide a validation mechanism over our proposed adaptation rule of the global adaptation model to guarantee the behavior correctness of the adaptation. Through this validation approach, computing the lifecycle of the global adaptation model can be avoided.

Ant-Colony Optimisation for Path Recommendation in Business Process Execution

In business process management, operational support concerns methods and tools to support users during the execution of business processes. One possible way of supporting users is to suggest the optimal way to complete the execution of a business process instance given the set of activities executed thus far and a notion of utility associated with the execution of possible remaining activities. This problem goes also under the label of process navigation. This paper proposes a novel technique to implement process navigation based on the innovative abstraction of business process models as a restricted class of directed hypergraphs, i.e. WF-hypergraphs. In our approach, workflow net process models are first transformed into WF-hypergraphs. Using this abstraction, finding the optimal way to complete a business process becomes a generalised hypergraph shortest path problem, which is NP-hard. To solve this problem, we propose a solution based on the ant-colony meta-heuristic specifically customised to the case of hypergraph traversal. The paper presents an experimental evaluation of the proposed optimisation heuristic and discusses how the proposed approach can be integrated into modern business process management systems.

Development of the course of process management of the enterprise with the use of free software

Goal. The goal of the search is to develop teaching methods of the process approach to enterprise management, which involves the execution of business process instances in a computer environment. The teaching methods are easy to use in any university, since during the course practical studies free software is used. It is available for installation via the Internet, which does not require payment or registration. The use of free software allows teaching the course partially or completely remotely. Students can get practical skills in working with the business process management system, as well as perform tasks outside the computer class by installing software on the home computer. The results of assignments can be uploaded to the campus or sent to the lecturer by an e-mail. Lecturers can also check completed tasks remotely, having installed the needed software to a laptop or a home computer. Materials and methods. During the development of the theoretical part of the course, a study was made of existing approaches to identify business processes in the enterprise, as well as techniques for automating business processes. The analysis of the concept of the process approach was carried out based on business processes performed in the computer environment of the enterprise, the component parts of the concept were identified and training materials were developed for studying each of the selected parts. The advantages of process automation and the training of the course were analyzed in such a way that during the training it was possible to show these advantages to students. In developing the practical part of the course, the authors’ experience in process automation of enterprises was used. In addition, the materials of the Association of Business Process Management Professionals (ABPMP) were published, as well as those published at various events of the association. Results. A methodology has been developed for training process management to students of financial and accounting specialties and students studying in the fields of Business Informatics, Applied Informatics (in Economics), and Automated Systems for Processing Information and Management. The methodology contains both theoretical and practical parts, as well as a set of tests and tasks used in tests and exams. The training course is aimed at forming among students three main competencies: developing business logic, setting up automatic generation of documents, organizing interaction with external data. For laboratory work, the free business process management system of RunaWFE is used. An approach was developed and tested in practice, within the framework of which the final control work of the course becomes part of the final qualifying work of the student. The course has been approved for several years in the National University of Science and Technology MISIS and Moscow State University of Economics, Statistics and Informatics (at present – Plekhanov Russian University of Economics). Classes for specific sections of the course were held in Ufa State Aviation Technical University, Bauman Moscow State Technical University, Financial University, Higher School of Economics (National Research University), Peoples’ Friendship University of Russia and Moscow Institute of Physics and Technology. Decisions that did not show good practical results were not included in the methodology. A text edition, containing theoretical material and laboratory workshop with a description of practical exercises and a set of control questions were published according to the course. Conclusion. The article demonstrates the experience of teaching students the process approach to enterprise management. This methodology allows training students to create business processes that implement solutions for various situations, arising in the business of enterprise. The gained knowledge will be useful for students who have received training in the follow-up work activity.

An ontology-based collaborative business service selection: contributing to automatic building of collaborative business process

With worldwide inter-enterprise collaboration and interoperability background, automatic collaborative business process deduction is a crucial researching subject. We have designed a methodology of deducing collaborative process by only collecting collaborative objectives and partners’ business services. The two key problems are (i) selecting corresponding business services for a set of collaborative objectives and (ii) ordering business services with serializations and parallelization. This paper aims to present a solution of business service selection and the following business process extraction. In order to solve the problem, we have defined a collaborative ontology, which contains numerous instances of business services and processes from the MIT process handbook. The collaborative ontology contains essential concepts in collaborative situations and process-deducing rules and algorithms. We provide a brief illustration of implementation within a SaaS toolkit called Mediator Modeling 2ool.

Time and activity sequence prediction of business process instances

The ability to know in advance the trend of running process instances, with respect to different features, such as the expected completion time, would allow business managers to timely counteract to undesired situations, in order to prevent losses. Therefore, the ability to accurately predict future features of running business process instances would be a very helpful aid when managing processes, especially under service level agreement constraints. However, making such accurate forecasts is not easy: many factors may influence the predicted features. Many approaches have been proposed to cope with this problem but, generally, they assume that the underlying process is stationary. However, in real cases this assumption is not always true. In this work we present new methods for predicting the remaining time of running cases. In particular we propose a method, assuming process stationarity, which achieves state-of-the-art performances and two other methods which are able to make predictions even with non-stationary processes. We also describe an approach able to predict the full sequence of activities that a running case is going to take. All these methods are extensively evaluated on different real case studies.

Analytical processing system of business processes, methods of exploration of business process definitions and instances

The article tackles the topics related to systems, methods and mechanisms for storing process definitions and instances for the purpose of their analytical processing - process warehouses and general architecture of such process warehouses, including tools allowing their extract, transform, load (ETL) as well as analytical processing and exploration of stored process definitions and instances - i.e. the process mining. Furthermore, the article describes a mechanism for storing processes, with the use of a network database, and one of the methods for researching into similarities of the processes through the tagging of generic processes.

A Robust and Versatile Multi-View Learning Framework for the Detection of Deviant Business Process Instances

Increasing attention has been paid to the detection and analysis of “deviant” instances of a business process that are connected with some kind of “hidden” undesired behavior (e.g. frauds and faults). In particular, several recent works faced the problem of inducing a binary classification model (here named deviance detection model ) that can discriminate between deviant traces and normal ones, based on a set of historical log traces (labeled as either deviant or normal). Current solutions rely on applying standard classifier-induction methods to a feature-based representation of the given traces, where the features include sequence-based patterns extracted from the corresponding sequences of activities. However, there is no consensus on which kinds of patterns are the most suitable for such a task. On the other hand, mixing multiple pattern families together may produce a heterogenous, redundant and sparse representation of the traces that likely leads to poor deviance detection models. In this paper, we propose an ensemble-learning method for solving this problem, where multiple base classifiers are trained on different feature-based views of the log (each obtained by mapping the traces onto a distinguished collection of patterns). A stacking procedure is used to combine the discovered base models into an overall probabilistic model that associates any new trace with an estimate of the probability that it reflects a deviant process instance. This helps the analyst prioritize the inspection of the cases that are more likely to be deviant. The method also takes advantage of all nonstructural data available in the log, and employs a resampling mechanism to deal with the rarity of deviances in the training log. It has been conceived as the core of a comprehensive framework for detecting and analyzing business process deviances. The framework supports the analyst to investigate suspect deviances, and provides some feedback to the learning method for improving the accuracy of the discovered deviance detection models. Tests on several real-life datasets proved the validity of the approach, as concerns its capability to discover an accurate deviance detection model, and to effectively exploit new (originally unlabeled) traces via active learning and self-training mechanisms.

Resource Management for Business Process Scheduling in the Presence of Availability Constraints

In the context of business process management, the resources required by business processes, such as workshop staff, manufacturing machines, etc., tend to follow certain availability patterns, due to maintenance cycles, work shifts and other factors. Such availability patterns heavily influence the efficiency and effectiveness of enterprise resource management. Most existing process scheduling and resource management approaches tend to tune the process structure to seek better resource utilisation, yet neglect the constraints on resource availability. In this article, we investigate the scheduling of business process instances in accordance with resource availability patterns, to find out how enterprise resources can be rationally and sufficiently used. Three heuristic-based planning strategies are proposed to maximise the process instance throughput together with another strategy based on a genetic algorithm. The performance of these strategies has been evaluated by conducting experiments of different settings and analysing the strategy characteristics.