A framework for optimisation based stochastic process discovery

Pierre Cry,András Horváth,Paolo Ballarini,Pascal Le Gall

2024-07-02

Abstract:Process mining is concerned with deriving formal models capable of reproducing the behaviour of a given organisational process by analysing observed executions collected in an event log. The elements of an event log are finite sequences (i.e., traces or words) of actions. Many effective algorithms have been introduced which issue a control flow model (commonly in Petri net form) aimed at reproducing, as precisely as possible, the language of the considered event log. However, given that identical executions can be observed several times, traces of an event log are associated with a frequency and, hence, an event log inherently yields also a stochastic language. By exploiting the trace frequencies contained in the event log, the stochastic extension of process mining, therefore, consists in deriving stochastic (Petri nets) models capable of reproducing the likelihood of the observed executions. In this paper, we introduce a novel stochastic process mining approach. Starting from a "standard" Petri net model mined through classical mining algorithms, we employ optimization to identify optimal weights for the transitions of the mined net so that the stochastic language issued by the stochastic interpretation of the mined net closely resembles that of the event log. The optimization is either based on the maximum likelihood principle or on the earth moving distance. Experiments on some popular real system logs show an improved accuracy w.r.t. to alternative approaches.

Databases

What problem does this paper attempt to address?

This paper mainly discusses an optimization-based framework for discovering stochastic process models that can accurately reproduce the execution behavior from event logs. Traditional process mining algorithms typically generate control flow models (such as Petri nets), but these models cannot capture the execution frequency information in event logs, i.e., the randomness of the process. Therefore, the paper proposes a new approach that starts with a standard Petri net mined through classical algorithms and optimizes the transition weights to make the model's stochastic language closer to the language of the event log.

Specifically, the paper first introduces the background and related concepts, including event logs, process mining, Petri nets, and stochastic process mining. Then, it describes a numerical optimization-based method that constructs a dedicated unfolded directed acyclic graph to calculate the probability of a given trace being generated by a stochastic Petri net, and adjusts the weights using the maximum likelihood principle or the earth mover's distance as the optimization objective. Experiments show that this method improves accuracy compared to existing methods.

Finally, the paper mentions related work, including other methods for discovering stochastic models from event logs, and outlines future directions for the proposed approach. The entire research focuses on leveraging frequency information from event logs to improve process models and better simulate the stochastic behavior of real-world business processes.