Resource allocation in stochastic networks


Stochastic network theory deals with analyzing congestion phenomena occurring in complex systems in a variety of settings. As concrete examples of these phenomena, one may think of traffic jams in road systems, delays of tasks in computer-communication systems, the build-up of products in intermediate stages of inventory/manufacturing systems, and the waiting time of customers calling to the call center of a retail company.


To mitigate the adverse effects of congestion as much as possible, optimal allocation of resources in these systems is crucial. For example, in computer-communication systems, enough resources need to be allocated to the right tasks to keep delays somewhat limited. Similarly, in manufacturing systems, enough processing machines need to be allocated to the correct products to keep the system running as efficiently as possible.  On the other hand, one does not want to deploy too many resources or machines out of economic considerations. The topic of the project thus mainly evolves around two generic questions. After addressing 1) how many resources are needed for a complex system to run efficiently, one needs to decide how to 2) divide these resources over the system to ensure optimal operation.

Supervisors Michel Mandjes (UvA) and Jan-Pieter Dorsman (UvA)
PhD Student Lucas van Kreveld
Location Universiteit van Amsterdam (UvA)