IT-architecture and coordination in transport chains

TNO - TRAIL research programme

prof.dr.ir. H.J. Sips (ITS/TUD, promotor)
ir.J. Rogier (TNO/TPD)
dr.ir. J.F.M. Tonino (ITS/TUD)
dr.C. Witteveen (ITS/TUD, project manager)

Interorganisational cooperation in multi-modal transportation systems

Interorganisational cooperation assumes an infrastructure which supports this cooperation and consequently supports means for information systems to exchange information according to a level agreed upon for the particular situation at hand.
To develop such an information system, in this project an algorithmic, multi-agent based approach is pursued to develop task allocation and task coordination processes that can be used by supply chain managers to coordinate the activities of several transportation actors.
In a task allocation process agents are allocated to parts of the transportation problem, while a task coordination process has to guarantee that the distributed problem solving activities are geared toward a solution for the complete problem. Clearly, these two processes are tightly interwoven: in order to guarantee a satisfactory solution, a poor task allocation policy will need an elaborate coordination process to achieve an acceptable solution and vice-versa.

Research Focus
The general research focus in this project is on the development of efficient and nearly optimal task allocation policies, as well as efficient coordination strategies, to solve complex logistic problems.

The research approach we will act upon is first to concentrate on task allocation policies and problem solving strategies based on them. We consider them as approximation algorithms used to allocate the tasks in a complex multi-agent setting to different agents in such a way that

    • the agents are able to solve their parts of the task almost independently from each other;
    • the resulting parts of the plan can be easily combined in a solution to the complete problem without requiring elaborate coordination;
    • the resulting assembled solution is nearly optimal.

As a result of this research project, a prototype of a decision-support tool will be developed to assist chain managers in task allocation processes.
Description and implementation of efficient coordination algorithms;
Reports about computational aspects.