HOW MACHINE LEARNING SUPPORTS PRIVACY CONCERNS
Machine learning (ML) techniques and artificial intelligence (AI) are often connected with privacy concerns. Prof. Pibernik has recently worked on a research project that uses these technologies in order to improve privacy preserving conditions of firms. What is the idea behind this project? Thinking about a collaborative planning approach, the involved firms have a keen interest that the transactional partner doesn’t have access to sensitive data. This is important because it would weaken their position in further negotiation processes. The goal of the research project was to develop tools that improve privacy preserving conditions – from a technical perspective – these tools are supporting the optimization process of data without disclosing the data in an encrypted data-base. Machine learning is a technique that is suited for this kind of question.
Within this research project it has been practiced with companies from the aerospace industry concerning the maintenance and service of engines. The collaborative planning approach affords a lot of data from different airlines – e.g. condition data, oil pressure etc. Obviously, merging the data across all participants brings the opportunity to create good forecasting mechanisms that improve the efficiency of the allocation process by knowing which engines need certain spare parts or which engines need for instance overall maintenance. But as the data of every single participant is very sensitive and confidential, they do not want it to be shared. The combination of machine learning and encrypted data-base technologies can create a framework that solves this problem and increases the efficiency of the transaction / integrated planning process – the approach can be practically applied to other industries as well.
DATA-DRIVEN ALLOCATION OF RESOURCES
In another current collaborative research project, Prof. Pibernik and his team are working together with Lufthansa Technik AG logistics services with the goal to develop a framework that allows predicting and planning capacity requirements in the field of maintenance repair and overall service (MRO) provision – the company provides its service to their own and other commercial airlines. The service is expansive and affords an efficient allocation of different spare parts and skilled labor resources.
The problem is, that it is very difficult to predict what type of spare part you need and what kind of maintenance and repair is needed given all logistics capacities. This process creates a large amount of data. The main goal is to consolidate that data and apply algorithms, machine learning and optimization techniques in order to determine and schedule their inbound logistics capacities and workload.
HOW DATA AND ALGORITHMS SUPPORT PHARMA-SUPPLY IN DEVELOPING COUNTRIES
The availability of essential medicines for the people in developing countries has improved throughout the last years. This is mainly due to the effort of a number of global health organizations. However, there is still an urgent need to bring down prices and to ensure that the people in these countries have access to the products. You can find a can of Coke in the most remote places in Africa, but not malaria medication. We commonly think that the main reason is that the people in developing countries cannot afford pharmaceutical products because they have no health insurance. This, however, is not the core problem. Very often, the reason is that efficient supply chains are non-existent.
Prof. Pibernik participated in multiple collaborative projects with various international institutions and start-ups with the objective to improve access to essential medicines. One of these projects is dedicated to improve the situation of local pharmacies in Kenya. The problem is that medical products are very expensive compared to other developing countries and very often are not available. One reason, why the products are expensive is that local pharmacies order in very small quantities that are too small to be placed directly at the manufacturer. They have to buy from various middlemen that charge comparably high prices. In order to improve the situation and reduce the ordering costs, the participating startup company equipped the pharmacies with tablet computers and a basic ERP-System. This technology allows pharmacies to manage their inventories more efficiently, to consolidate orders across multiple pharmacies and to order directly from the manufacturer or a wholesaler. The team of Prof. Pibernik has been working on forecasting an inventory algorithms that consolidated orders across multiple pharmacies with the goal to find an optimal mechanism to compete for big orders that leads to lower prices and increases the availability of the products (more information about the projects: https://www.wiwi.uni-wuerzburg.de/lehrstuhl/bwl11/research/multi-supplier-sourcing-strategies-for-global-health-products/ ).
The mentioned research projects are focusing on different industries and topics, but they are based on similar techniques (optimization, machine learning etc.) with the goal to improve parts of the supply chain.
MIT NETWORK IN LOGISTICS
Prof. Pibernik has been involved in the MIT’s (Massachusetts Institute of Technology) Global SCALE network for 15 years. This network consists of six centers of excellence in research and education on four continents. At each center, there are very active and acknowledged researchers that work on leading-edge projects and provide top-of-the-world education/masters programs in the field of supply chain management (Link https://scm.mit.edu/). The Gobal SCALE Network was founded by Prof. Dr. Yossi Sheffi, one of the most prominent academics in logistics and supply chain management and entrepreneur. Prof. Scheffi is director of the MIT Center for Transportation and Logistics (MIT CTL), professor of Civil and Environmental Engineering at the Massachusetts Institute of Technology. Prof. Pibernik is an Adjunct Professor at the Global SCALE center in in Zaragoza, Spain, that was implemented in 2003 and a Visiting Professor at the Malaysia Institute of Supply Chain Innovation that was inaugurated in 2011.
For more details, listen to the full podcast and leave me a reedback and rating!