Dynamic Principal-Agent Framework for Optimal Contract Design in Railway Networks


Staff Mentor:

prof.dr.ir. B. De Schutter (Bart)


Other Mentor(s):

Z. Su

Keywords:

Railway networks; Transportation and infrastructure; Optimization-based control

Description:

Since the privatization of Dutch railways, the maintenance of the railway infrastructure is performed by private contractors, whose short-term objectives within the contracted period are not fully aligned with the long-term objectives of the infrastructure manager ProRail. The current contract between ProRail and the contractors is a performance-based contract, which can be represented by the figure below. The interaction between ProRail and a contractor under such a performance-based contract can be described by a principal-agent model, in which ProRail is modelled as the principal and the contractor modelled as an agent. ProRail delegates a maintenance task to the contractor, who executes this task with an effort level suggested by ProRail. After completion of the maintenance task, ProRail pays the contractor according to the measured performance output, which is also influenced by random environmental factors, i.e. bad weather, accidents, etc.

At first, the MSc student should conduct a literature review on mechanism design to get an overview of the field and to identify possible research directions. The aim is to develop a new methodology for optimal contract design applied to maintenance of railway networks using a dynamic principal-agent framework. Starting from a static single-principal-single-agent model with uncertainty, a dynamic model considering track degradation should be constructed. Based on the developed game-theoretic model for the contractual relationship between ProRail and the contractor(s), a general methodology for optimal contract design can be worked out, with dynamic programming as a possible solution approach.

If you are interested in this project, please contact z.su-1 AT tudelft.nl


© Copyright Delft Center for Systems and Control, Delft University of Technology, 2017.