TU Delft Algorithmics
Mathijs M. de Weerdt
Delft University of Technology Projects ALG Group
EWI ALG Mathijs de WeerdtProjects
eng
 
 
 
 
 
 
 
 
 
Coordinated multiagent planning
VENI
My main project is my VENI project on Coordinated multiagent planning financed by NWO/STW (2005-2008). The following is an introduction to this research.
There is a growing interest towards methods that support the coordination of activities of different companies. For example, if trucking companies are able to coordinate their truck planning, they can share empty rides to reduce their operational costs significantly. Since such coordination processes are computationally intensive and need to be highly adaptive, human-based support for coordination is simply not feasible. Consequently, ICT-support for inter-organizational coordination is needed.
This proposal aims to develop an ICT-based framework to support inter-organizational coordination. In particular, it aims to develop a system to support autonomous planners (i.e., the planning departments of the companies involved) in coordinating their planned activities to enhance their individual and common efficiency.
In this research proposal special attention will be given to privacy and reusability aspects of coordination. This requires a new approach to coordinated planning, since current theories on coordination and planning assume that all parties are willing to share all information. In practice, however, (competing) companies will often refuse to share knowledge. This leads to the requirement that each agent should be able to decide which information to share and when to participate. Furthermore, companies would like to continue using their own planning systems. Therefore, the theoretical kernel of this proposal is the development of a conceptual framework and algorithms for coordinated planning that enable participating companies (i) to maintain their own level of autonomy and privacy and (ii) to reuse their existing planning tools.
The results of this research will be validated empirically. We aim to show that not only the operational costs of the companies involved will be reduced, but that also the customers can benefit: we expect that by coordination companies are able to offer their services more flexibly, more reliably (e.g., by using other companies resources), and for less costs.
This project started January 1st 2005 and will take until December 31st 2008.
See my publications for publicly available scientific results, or log in here for all documents related to this project.
Other projects
I am involved in
I am also involved in Sicco Verwer's motor management project on inductive inference of (timed) automata.
My work is part of the CABS program.
                                                                                                                                                                                                                                                                             
CABS
Collective Agent-Based Systems
Top of the page