In the context of the “Process Mining in Logistics” research project between Vanderlande Industries, we are offering multiple Master projects aimed at laying the foundations for a “Thinking Assistant” for large-scale material handling systems. Such a “Thinking Assistant” shall support engineers and operators in faster identifying problems and root-causes, predicting possible problems, and proposing solutions that resolve problems and optimize system usage.
To realize this aim, the project aims at integrating process mining, knowledge graphs, and machine learning techniques over very large data (logistics processes are a factor 10-100 larger than other processes). We offer several topics that build on a rich stack of recently developed advanced techniques for data preparation, process mining, and process prediction.
We are offering multiple projects, three concrete topics are listed below. Other topics can be identified.
Predictive Model(s) For Operational Planning for Customers of Automated Material Handling Solutions
Vanderlande customers using the automated Material Handling Solutions in Warehouses and Airports have two main needs namely to reduce the number of late bags or late orders and optimize the number of employees like operators that are required to operate the system. The optimization of these KPIs is key in determining the Total Cost of Ownership for the customer. The goal of this project is to build a solution that can predict the number of resources required to ensure the number of late bags or late orders are reduced given the historic data of plans and evaluation measures for the plans. This solution can be built in multiple ways like updating existing simulation models to replay data from actual system and extending it to forecast the future, building reinforcement learning algorithms or building machine learning models that could predict the required parameters.
The objectives of this graduation topic is to:
- Determine the resources that has an impact on the late bags or orders and impact the Total Cost of Ownership
- Evaluate the different technologies that can make better plans than existing plans and suggest the best technology to use for this optimization problem
- Build a prediction model based on the best technology to build the solution to make a better plan
Extend Knowledge Graphs to Predict System and Item Behavior
As a part of the ongoing research, a knowledge-graph is implemented that can describe the process related to the airport baggage handling systems, relate it to the physical system layout, show the interactions between the different process steps, the durations between different process steps and some of the characteristics of the bags. The knowledge-graph can provide answers on where there were delays in system for a bag and relate to the system and bag characteristics around this delay. The goal of this graduation project is to extend the knowledge-graph to include more information, aggregate information and use it for predictions.
The objectives of this graduation project are to:
- Extend existing knowledge to include how unexpected and undesired behaviour like delays propagate through the system
- Include data sources that can provide additional context information
- Aggregate information to provide overview at a higher level of processes
- Build a proof-of-concept and evaluate the feasibility to include this in existing products
- Extend the knowledge graph to predict unexpected and undesired behavior
Data Analysis and Visualization Techniques for Multi-Item Processes
Vanderlande offers warehouse solution projects like STOREPICK. In these solutions the items move through the system as different item types. Items arrive to the Warehouse as a part of pallets, in the warehouse the items are made into smaller units and placed in trays that transport the items in the system, and finally the items are put together to fulfil an order onto a roll cage or a pallet. The aim of this graduation project is to find the best representation of this information that can used to do a root cause analysis when an order is not fulfilled on time. The items of the same category for example all Coca Cola have the same item identifier which is the SKU. The pallets, roll cages and trays have their own identifiers. Any tray with a given SKU can be used to fulfil an order that required that particular SKU.
The objective of this graduation topics is to explore techniques that can:
- Represent the warehouse process such that if an order is delayed it can be drilled down to find all the trays that fulfilled the order and the pallets from where the items entered the warehouse
- Represent the system characterics like system errors such that it can relate to the warehouse process and provide more context to delay in orders
- Represent the required SKUs for an order and relation to existing trays with these SKUs to find if there are better solutions to fulfil an order
Vanderlande is the global leader for value-added logistic process automation at airports and in the parcel market, as well as a leading supplier of process automation solutions for warehouses. In this world of technology, we believe in people – especially those who are totally dedicated, customer-driven and keen to continue learning throughout their career. Some figures:
- Vanderlande’s baggage handling systems move 3.7 billion pieces of luggage around the world per year.
- Our systems are active in 600 airports including 13 of the world’s top 20.
- More than 39 million parcels are sorted by its systems every day, which have been installed for the world’s leading parcel companies.
- More than 206 projects in 105 countries
- Many of the largest global e-commerce players and distribution firms have confidence in Vanderlande’s efficient and reliable solutions.
Vanderlande focuses on the optimization of its customers’ business processes and competitive positions. Through close cooperation, we strive for the improvement of our customers’ operational activities and the expansion of their logistical achievements.
The projects are offered by the Digital Service Factory. The Digital Service Factory (DSF) is a new department being established within the Technology department (former R&D). The DSF develops data-driven digital services – such as a customer portal, connected systems, performance reporting and predictive maintenance – in close cooperation with the Vanderlande Business Units (BU’s) and its customers. The department is organized around dedicated multidisciplinary product teams who work closely together with other teams from Technology, ICT, BU’s and external partners. Key in this are service design thinking, agile way of working and a relentless focus on data and insights.
Previous Master projects
- A Forecasting Framework for Recirculation in Baggage Handling Systems, van der Sanden, G., https://research.tue.nl/en/studentTheses/a-forecasting-framework-for-recirculation-in-baggage-handling-sys
- Route and Performance Analysis of Multi-Item Processes in Warehouse Automation Systems, Homvanish, S., https://research.tue.nl/en/studentTheses/route-and-performance-analysis-of-multi-item-processes-in-warehou
- Context-Aware Performance Deviation Analysis for Logistics Systems, Boender, J., https://research.tue.nl/en/studentTheses/context-aware-performance-deviation-analysis-for-logistics-system
- Performance-aware conformance checking on material handling systems: visualization methodology of performance of routes in a baggage handling system Turu Pi, A. (Author). 30 Sep 2019 https://research.tue.nl/en/studentTheses/performance-aware-conformance-checking-on-material-handling-syste
- Outlier detection in event logs of material handling system. Köroglu, Ö. (Author). 30 Sep 2019. https://research.tue.nl/en/studentTheses/outlier-detection-in-event-logs-of-material-handling-system
- Process mining for systems with automated batching: an exploratory study on new process mining grounds Verhaegh, K. W. (Author). 24 Sep 2018. https://research.tue.nl/en/studentTheses/process-mining-for-systems-with-automated-batching
See https://pa.win.tue.nl/master-projects/ for the procedure
Dirk Fahland firstname.lastname@example.org