Research work
The proposed CD Laboratory addresses important planning and scheduling problems that arise in industry. The problems are challenging and require robust search techniques. Our aim is to develop fundamentally new solution techniques for these tremendously complex tasks and to contribute significantly to solutions techniques for other NP-hard problems. The main research aims are:
- Developing new techniques that provide high-quality solutions for new large-scale and complex planning and scheduling problems that appear in industry and other areas. Which solving paradigms (metaheuristics, CP, ASP, SAT, IP …) are most appropriate for such problems?
- Further automation of the process of planning and scheduling by providing methods that are adaptable and can be applied to different problems.
- Contributing to automated algorithm selection by designing new and problem independent features.
- Designing methodologies for automated generation of heuristics and search components.
- Exploiting machine learning in the process of automation of techniques for planning and scheduling.
The CD Laboratory includes three modules, each featuring a single industrial partner:
Module I: Adaptive techniques in personnel planning and scheduling. Industrial partner: XIMES.
Module II: Production planning and scheduling problems Industrial partner: MCP
Module III: Intelligent optimization for Electromagnetic Compatibility Laboratory Industrial partner: Bosch
Each module investigates the use of a particular problem-solving paradigm for problems from our industrial partners, with each module assigned to a single industrial partner. The work has the character of basic research and we aim throughout at developing methods and solutions that can be widely applied. Much of the work in each module will is performed in close cooperation with the industrial partners, who provide real-life problems, test our solutions and contribute ideas to help with the solution of their problems.