Automated evolutionary restructuring of workflows to minimise errors via stochastic model checking

This paper presents a framework for the automated restructuring of workflows that allows one to minimise the impact of errors on a production workflow. The framework allows for the modelling of workflows by means of a formalised subset of the Business Process Modelling and Notation (BPMN) language, a well-established visual language for modelling workflows in a business context. The framework’s modelling language is extended to include the tracking of real-valued quantities associated with the process (such as time, cost, temperature). In addition, this language also allows for an intention preserving stochastic semantics able to model both probabilisticor non-deterministic branching behaviour. We further extend this formalism to allow for the introduction of error states which allow for both fail-stop behaviour and continued system execution. We explore the practical utility of this approach by means of a case study from the food industry. Through this case study we explore the extent to which the risk of production faults can be reduced and the impact of these can be minimised, primarily through restructuring of the production workflows. This approach is fully automated and only the modelling of the production workflows and the expression of the goals require manual input.