A recommendation system for predicting risks across multiple business process instances
Introduction
A process-related risk measures the likelihood and the severity that a negative outcome, also called fault, will impact on the process objectives [1]. Failing to address process-related risks can result in substantial financial and reputational consequences, potentially threatening an organization's existence. Take for example the case of Société Générale, which went bankrupt after a €4.9B loss due to fraud.
Legislative initiatives like Basel II [2] and the Sarbanes–Oxley Act1 reflect the need to better manage business process risks. In line with these initiatives, organizations have started to incorporate process risks as a distinct view in their operational management, with the aim to effectively control such risks. However, to date there is little guidance as to how this can be concretely achieved.
As part of an end-to-end approach for risk-aware Business Process Management (BPM), in [3], [4], [5] we proposed several techniques to model risks in executable business process models, detect them as early as possible during process execution, and support process administrators in mitigating these risks by applying changes to the running process instances. However, the limitation of these efforts is that risks are not prevented, but rather acted upon when their likelihood exceeds a tolerance threshold. For example, a mitigation action may entail skipping some tasks when the process instance is very likely to exceed the defined maximum cycle time. While effective, mitigation comes at the cost of modifying the process instance, often by skipping tasks or rolling back previously-executed tasks, which may not always be acceptable. Moreover, we have shown that it is not always possible to mitigate all process risks [4]. For example, rolling back a task for the sake of mitigating a risk of cost overrun, may not allow the full recovery of the costs incurred in the execution of that task.
To address these limitations we propose a recommendation system that supports process participants in taking risk-informed decisions, with the aim to reduce process risks preemptively. A process participant takes a decision whenever they have to choose the next task to execute out of those assigned to them at a given process state, or via the data they enter in a user form. This input from the participant may influence the risk of a process fault to occur. For each such input, the technique returns a risk prediction in terms of the likelihood and severity that a fault will occur if the process instance is carried out using that input. This prediction is obtained via decision trees which are trained using historical process data such as process variables, resources, task durations and frequencies. The historical data of a process is observed using decision trees which are built from the execution logs of the process, as recorded by the IT systems of an organization.
This way, the participant can take a risk-informed decision as to task to execute next, or can learn the predicted risk of submitting a form with particular data. If the instance is subjected to multiple potential faults, the predictor can return the weighted sum of all fault likelihoods and severities, as well as the individual figures for each fault. The weight of each fault can be determined based on the severity of the fault's impact on the process objectives.
The above technique only provides “local” risk predictions, i.e. predictions relative to a specific process instance. In reality, however, multiple instances of (different) business processes may be executed at any time. Thus, we need to find a risk prediction for a specific process instance that does not affect the prediction for other instances. The interplay between risks relative to different instances can be caused by the sharing of the same pool of process participants: two instances may require the same scarce resource. In this setting, a sub-optimal distribution of process participants to the set of tasks to be executed may result in a risk increase (e.g. overtime or cost overrun risk). To solve this problem, we equipped our recommendation system with a second technique, based on integer linear programming, which takes input from the risk prediction technique, to find an optimal distribution of process participants to tasks. By optimal distribution we mean one that minimizes the overall execution time (i.e. the time taken to complete all running instances) while minimizing the overall level of risk. This distribution is used by the recommendation system to suggest to process participants the next task to perform.
We operationalized our recommendation system on top of the YAWL BPM system by extending an existing YAWL plug-in and by implementing two new custom YAWL services. This implementation prompts process participants with risk predictions upon filling out a form or for each task that can be executed. We then evaluated the effectiveness of our recommendation system by conducting experiments using a claim handling process in use at a large insurance company. With input from a team of risk analysts from the company, this process has been extensively simulated on the basis of a log recording one year of completed instances of this process. The recommendations provided by our recommendation system significantly reduced the number and severity of faults in a simulation of a real life scenario, compared to the process executed by the company as reflected by the event data. Further, the results show that it is feasible to predict risks across multiple process instances without impacting on the execution performance of the BPM system.
The remainder of this paper is organized as follows. Section 2 discusses related work. Section 3 contextualizes the recommendation system within our approach for managing process-related risks, while Section 4 presents the YAWL language as part of a running example. Next, Section 5 defines the notions of event logs and faults which are required to explain our techniques. Section 6 describes the technique for predicting risks in a single process instance while Section 7 extends this technique to the realm of multiple process instances running concurrently. 8 Implementation, 9 Evaluation discuss the implementation and evaluation of the recommendation system, respectively. Finally, Section 10 concludes the paper. A provides the formal definition of a YAWL specification, the algorithms to generate a prediction function, and technical proofs of two lemmas presented in Section 7.
Section snippets
Related work
The approach presented in this paper is related to work on risk prediction, job scheduling, operational support and work-item distribution for business processes. In this section we review the state of the art in these fields to motivate the need for our approach.
Risk framework
In this section we elaborate on the type of process-related risks that we can address and on the basis of this, we illustrate an overarching approach for managing process-related risks within which the contribution of this paper fits.
YAWL specification and running example
We developed our technique on top of the YAWL language [41] for several reasons. First, this language is very expressive as it provides a comprehensive support for the workflow patterns,2 patterns covering all main process perspectives such as control-flow, data-flow, resources, and exceptions. Further, it is an executable language supported by an open-source BPM system, namely the YAWL System. This system is based on a service-oriented architecture, which facilitates
Event logs and fault severity
The execution of completed and running process instances can be stored in an event log: Definition 1 Event log Let T and V be a set of tasks and variables, respectively. Let U be the set of values that can be assigned to variables. Let R be the set of resources that are potentially involved during the execution. Let D be the universe of timestamps. Let Φ be the set of all partial functions V
Risk estimation
We aim to provide work-item recommendation to minimize the risk corresponding to the highest product of fault severity and likelihood. For this purpose, it is necessary to predict the most likely fault severity associated with continuing the execution of a process instance for each enabled task. The problem of providing such a prediction can be translated into the problem of finding the best estimator of a function. Definition 2 Function estimator Let X1, …, Xn be n finite or infinite domains. Let Y be a finite domain. Let f : X1 × X
Multi-instance work-item distribution
With the technique presented so far, each resource is given local risk advice as to what work item to perform next, i.e. a resource is suggested to perform the work item with the lowest overall risk for that combination of process instance and resource, without looking at other resources that may be assigned work items within the same instance or in other instances running concurrently. Clearly, such a local work-item distribution is not optimal, since work items have to compete for resources
Implementation
We operationalized our recommendation system on top of the YAWL BPM system, by extending an existing YAWL plug-in and by implementing two new custom YAWL services. This way we realized a risk-aware BPM system supporting multi-instance work distribution and form filling-out.
The intent of our recommendation system is to “drive” participants during the execution of process instances. This goal can be achieved if participants can easily understand the suggestions proposed by our tool. For this we
Evaluation
We evaluated our recommendation system using the claim handling process and related event data, of a large insurance company kept under condition of anonymity. The event data recording about one year of completed instances (total: 1065 traces) was used as a benchmark for our evaluation. The claim handling process, modeled in Fig. 6, starts when a new claim is received from a customer. Upon receipt of a claim, a file review is conducted in order to assess the claim, then the customer is
Conclusion
This paper proposes a recommendation system that allows users to take risk-informed decisions when partaking in multiple process instances running concurrently. Using historical information extracted from process execution logs, for each state of a process instance where input is required from a process participant, the recommendation system determines the risk that a fault (or set of faults) will occur if the participant's input is going to be used to carry on the process instance. This input
Acknowledgments
This research is partly funded by the ARC Discovery Project “Risk-aware Business Process Management” (DP110100091). NICTA is funded by the Australian Government as represented by the Department of Broadband, Communications and the Digital Economy and the Australian Research Council through the ICT Centre of Excellence program.
Raffaele Conforti is a post-doctoral research fellow in the Information Systems School in the Science and Engineering Faculty, Queensland University of Technology, Brisbane, Australia. He received his PhD from Queensland University of Technology in 2014. He is conducting research in the area of business process management. His research focuses on incorporating aspects of risk management into business process management systems. In particular he focuses on the identification of techniques for
References (45)
- et al.
Real-time risk monitoring in business processes: a sensor-based approach
Journal of Systems and Software
(2013) A reinforcement learning approach to dynamic resource allocation
Engineering Applications of Artificial Intelligence
(2007)Resource allocation via dynamic programming in activity networks
European Journal of Operational Research
(1993)Utility-based decision support system for schedule optimization
Decision Support Systems
(2008)- et al.
Time prediction based on process mining
Information Systems
(2011) A work system view of DSS in its fourth decade
Decision Support Systems
(2004)- et al.
Visual support for work assignment in process-aware information systems: framework formalisation and implementation
Decision Support Systems
(2012) - et al.
Mining association rules to support resource allocation in business process management
Expert Systems with Applications
(2011) - et al.
A semi-automatic approach for workflow staff assignment
Computers in Industry
(2008) Standard AS/NZS ISO 31000
(2009)
Basel II — International Convergence of Capital Measurement and Capital Standards
History-aware, real-time risk detection in business processes
Automated risk mitigation in business processes
OCTAVE criteria, version 2.0
The Use of the CCTA Risk Analysis and Management Methodology CRAMM in Health Information Systems
Model-driven Risk Analysis — The CORAS Approach
Current Research in Risk-aware Business Process Management―Overview, Comparison, and Gap Analysis
Communications of the Association for Information Systems:
Predicting deadline transgressions using event logs
Root cause analysis with enriched process logs
Introduction to Sequencing and Scheduling
A reinforcement learning approach to job-shop scheduling
Dynamic work distribution in workflow management systems: how to balance quality and performance
Journal of Management Information Systems
Cited by (137)
Combining an LNS-based approach and organizational mining for the Resource Replacement Problem
2024, Computers and Operations ResearchA unified model of supply risk mitigation
2023, Computers and Industrial EngineeringA proposed guidance approach for BP performance improvement
2023, Procedia Computer ScienceEncoding resource experience for predictive process monitoring
2022, Decision Support Systems
Raffaele Conforti is a post-doctoral research fellow in the Information Systems School in the Science and Engineering Faculty, Queensland University of Technology, Brisbane, Australia. He received his PhD from Queensland University of Technology in 2014. He is conducting research in the area of business process management. His research focuses on incorporating aspects of risk management into business process management systems. In particular he focuses on the identification of techniques for automatic detection, prevention and mitigation of risks that may eventuate during the execution of business processes.
Massimiliano de Leoni is an Assistant Professor of Information Systems at the Technische Universiteit Eindhoven (TU/e), The Netherlands. In 2009, he earned a Ph.D. in Computer Engineering at SAPIENZA - University of Rome (Italy) discussing a dissertation on “Adaptive Process Management in Highly Dynamic and Pervasive Scenarios”. He has been guest research fellow at Queensland University of Technology, Vienna University of Economics and Business and University of Naples. His research interests are in the area of Process-aware Information Systems and Business Process Management, predominantly focusing on multi-perspective process mining, process-aware decision support systems and visualization techniques for business process management and analysis.
Marcello La Rosa is associate professor of business process management and academic director for corporate engagements for the Information Systems School at the Queensland University of Technology, Brisbane, Australia. He researches on process consolidation, mining, configuration and automation. He received his PhD degree from Queensland University of Technology in 2009.
Wil van der Aalst is a full professor of Information Systems at the Technische Universiteit Eindhoven (TU/e). He is also an adjunct professor at Queensland University of Technology (QUT). His research interests include workflow management, process mining, Petri nets, business process management, process modeling, and process analysis. He is an elected member of the Royal Holland Society of Sciences and Humanities (Koninklijke Hollandsche Maatschappij der Wetenschappen) and the Academy of Europe (Academia Europaea).
Arthur ter Hofstede is a Professor in the Information Systems School in the Science and Engineering Faculty, Queensland University of Technology, Brisbane, Australia, and is Head of the Business Process Management Discipline. He is also a Professor in the Information Systems Group at Eindhoven University of Technology, Eindhoven, The Netherlands. His main research interests lie in the areas of business process automation and process mining.