Application and Improvement of Computational Intelligence Algorithms for Modeling and Optimization of Complex Problems in Supply Chain Planning and Epidemics Forecasting
Description:
Planning the activities in a large supply and distribution network is a highly complex task involving a large number of actors, deciding about a multitude of parameters like production and transportation volumes or inventory levels. Current methods being applied are based on classical methods of operations research or different meta-heuristics, resulting very often in not acceptable run times. Besides the sheer complexity of the planning task the involvement of independent actors requires methods of decentralized planning. From the computational side mainly agent-based methods are being applied to cover this aspect. The same complexity is found in epidemiological modeling of viral disease outbreaks, as hidden relations among actors have to be understood for usable forecast to be produced.
Regarding the optimization of such complex problems new computational intelligence algorithms (e.g. fish school search – FSS) are promising an efficient computation while on the same time leading to results of high quality. The integration with agent-based social simulation (ABSS) offers new possibilities in describing and understanding the dynamics of, on the one hand, the supply chain planning process between independent actors and, on the other hand, the parameters that control an epidemic spread. Thus we foresee that the hybridization of a swarm intelligence technique and an agent-based technique will be able to tackle the difficulties posed by the selected application problems as they blend fast exploratory capabilities of large search spaces (by FSS) and meaningful representations of spatiality such as neighboring and locality (given by ABSS).
In this context the following scientific objectives are approached through the project:
- Application of new modeling and optimization techniques within the areas as diverse as supply chain planning and epidemic spread modeling
- Integration of these techniques with the existing state of the art planning and simulation methods in supply chain planning and epidemic spread modeling
- Adaptation and extension of the methods from computational intelligence
- Evaluation of the methods regarding the solution quality and efficiency
- Investigation of further application areas (e.g. logistics in public health) and possibilities for industrial deployment
| Project status | finished |
| Project time | 09/2013-08/2016 |
| Group | Chair for Information Systems and Supply Chain Management (Prof. Dr.-Ing. Bernd Hellingrath) |
| Funding source | BMBF |
| Project number | 01DN13033 |
| Keywords | Computational Intelligence; Supply Chains; Epidemics |