On the 15th of April, 10.00 am, IW 5.57, Lu Xia will talk about applying Process mining to discover and optimise retail processes:

In this seminar, I will cover a contribution presented at the 7th International Conference on Process Mining (ICPM 2025), 2nd Workshop on Process Mining for Sustainability (PM4S).

Food waste represents a significant global challenge, with a substantial portion occurring at the retail level due to the misalignment between supply strategies and fluctuating customer demand. This seminar presents a novel, data-driven framework that integrates Object-Centric Process Mining (OCPM) with Stochastic Process Analysis to optimize inventory management and reduce waste.

We first demonstrate how to discover a stochastic process, a Continuous-Time Markov Chain (CTMC), directly from grocery store event logs. Unlike traditional methods, our approach models the inventory level as a set of dynamic states where transitions are triggered by real-time customer purchasing behaviour. To ensure the mathematical rigor of the exponential arrival assumption in CTMCs, we developed an automated statistical pipeline and conducted large-scale distribution fitting for over 227 unique product identifiers. This base model is then enhanced with supply activities to conduct "what-if" analyses, allowing retailers to quantify the trade-off between product shortages (under supply) and food waste (over supply).

Our results demonstrate that by leveraging event data, retailers can move beyond static replenishment rules toward dynamic strategies that significantly lower environmental impact while maintaining high service levels.

On the 28th of November, 11.00 am,  IW 5.57, we are planning to run the second process where Dr. Adam Banham will tell us about his research of the stochastic Petri nets and external data streams! An exciting topic on how external data sources influence a process model. Please join us in person or online! 

This is what Adam says about his research: 

In this seminar, I will cover a recent contribution presented at the International Conference on Application and Theory of Petri Nets and Concurrency (2025). I will present the paper's contributions in a visual manner for all audiences, not just enthusiasts of Petri nets. Please see the abstract below for a high-level summary. Process mining, when applied to data stored in a business's information systems, provides insights into the internal performance of its processes. These insights reveal how the behaviour of processes impacts businesses, and can inform planning for various future scenarios by anticipating how processes will perform in these scenarios.

However, these scenarios may be influenced by the context of the process, i.e., external data streams (exogenous data) to the process. In these cases, typical process discovery techniques can produce process models that describe the activities that could occur next in a given state, but struggle to express the effect of external influences on the likelihood of these activities occurring. This contribution presented an extension of stochastic labelled Petri nets and a discovery technique for our new modelling formalism. The proposed formalism can be used to quantify whether the firing likelihood of a transition is influenced by exogenous data when replaying historical process executions over the net. Comparing our approach with existing stochastic techniques using several publicly available event logs showed that our approach can outperform existing data-aware techniques in unstructured processes.

https://adambanham.io/papers/2025/Petri_Nets_2025___Exogenous_Influences_on_Processes.pdf

We are happy to share that on the 8th of October (1-2 pm, Adelaide time) we will run the first Process mining seminar!

Where Ethan Johnson will talk about the discovery of Social online processes using Process mining.

Please join us in 5.57, IW, North Terrace Campus. 

Title: Process Mining for Social Networks: Uncovering AI and Human Behaviour Online

Abstract: The rapid growth of social media presents a unique opportunity to study coordinated agent behaviour in an unfiltered environment. Online processes often exhibit complex structures that reflect the nature of the user behaviour, whether it is authentic and genuine, or part of a coordinated effort by malicious agents to spread misinformation and disinformation. Detection of AI-generated content can be extremely challenging due to the high quality of large language model-generated text.Therefore, approaches that use metadata like post timings are required to effectively detect coordinated AI-driven campaigns. Existing work that models the spread of information online is limited in its ability to represent different control flows that occur within the network in practice. Process mining offers techniques for the discovery of process models with different routing constructs and are yet to be applied to social networks. We propose to leverage process mining methods for the discovery of AI and human agent behaviour within social networks. By applying process mining techniques to real-world Twitter (now X) event data, we demonstrate how the structural and behavioural properties of discovered process models can reveal coordinated AI and human behaviours online.