Refine
Document Type
- Article (1)
- Part of a Book (1)
Language
- English (2)
Publication reviewed
- begutachtet (2)
Keywords
- Carsharing (1)
- Electric vehicle (1)
- Generative adversarial network (1)
- Machine learning (1)
- Shared mobility (1)
- Synthetic data (1)
Enabling Flexible Laboratory Processes – Designing the Laboratory Information System of the Future
(2018)
Situation faced: Recent developments in the medical and industrial laboratory market have increased the need for highly flexible laboratory processes. This pressure results from new requirements that have accompanied the internationalization of laboratories and the digitalization of paper-based, bureaucratic work practices. The execution of laboratory processes is supported by laboratory information systems (LISs), which handle the control and information flow of incoming orders end-to-end. State-of-the-art LISs do not feature sufficient flexibility-to-use and flexibility-to-change capabilities. To prepare medical and industrial laboratories for the challenges ahead, LISs require more advanced flexibility capabilities that meet the need for flexibility in complex laboratory processes.
(b)
Action taken: To address the challenges of medical and industrial laboratories, MELOS, a leading German LIS provider, and the Project Group BISE of the Fraunhofer FIT conducted the LIS4FUTURE project. The project team compiled requirements on the flexibility of laboratory processes and derived corresponding requirements for the LIS’s flexibility-to-use and flexibility-to-change. The lack of configuration capabilities and modularity across all layers of the software architecture was identified as a major inhibitor of flexible laboratory processes. Following an agile development process and grounded on extant knowledge, the project team developed the LIS4FUTURE demonstrator, a process-aware LIS with a modular architecture and a rule-based configuration mechanism.
(c)
Results achieved: Based on identified requirements, the project team iteratively developed and evaluated the modular architecture and the rule-based configuration mechanism as part of the development of the LIS4FUTURE demonstrator. The modular architecture allows for the complete replacement of process steps at build time, while the rule-based configuration mechanism makes it possible to meet the ever-increasing demands for flexibility at runtime. The LIS4FUTURE demonstrator, which shows the applicability of the developed concepts in real-world scenarios, will help MELOS develop an innovative release of their LIS.
(d)
Lessons learned: During the LIS4FUTURE project, the project team learned that: (1) advanced flexibility-to-use and flexibility-to-change IS capabilities are needed to prepare for flexibility demands on the process level; (2) radical redesign of existing processes and systems should be preferred over incremental improvement in order to tap the disruptive potential of innovation opportunities; (3) the LIS architecture must be aligned with the process paradigm if it is to be flexible; (4) discussions among academics and practitioners are more effective if they are based on running prototypes rather than on theoretical concepts; and (5) project results improve if project team members work a substantial fraction of their time at the same location.
Carsharing is an integral part of the transformation toward flexible and sustainable mobility. New carsharing programs are entering the market to challenge large operators by offering innovative services. This study investigates the use of generative machine learning models for creating synthetic data to support carsharing decision–making when data access is limited. To this end, it explores the evaluation, selection, and implementation of leading-edge methods, such as generative adversarial networks (GANs) and variational autoencoders (VAEs), to generate synthetic tabular transaction data of carsharing trips. The study analyzes usage data of an emerging carsharing program that is expanding its services to include free-floating electric vehicles (EVs). The results show that augmenting real training data with synthetic samples improves predictive modeling of upcoming trips by up to 4.63%. These results support carsharing researchers and practitioners in generating and leveraging synthetic mobility data to develop solutions to real-world decision support problems in carsharing.