Scalable Resource Allocation Methods

Témavezető (TMIT): 
Doktori iskola: 
Villamosmérnöki Tudományok Doktori Iskola
Kiírás dátuma: 
2017. 01. 03
Érvényesség: 
inaktív
Téma leírása: 

Research objectives: Resource allocation mechanism has been long studied in the literature. Both centralized and distributed solutions havebeen investigated. One of the modeling tools for resource allocation in communication networks is the processor sharing approaches. Many variants of this discipline have already been investigated in several papers. The discriminative type models have attracted attention in modeling different levels of quality of service in terms of the allocated bandwidths and expected service times.

These models are to be extended and further investigated in order to take into account the technological limits on access to the resources. For example, if classes of traffic flows are going to share communication capacity in a weighted proportional manner, the access rate limitations can prohibit to provide fully the proportional allocation. In this case the mixed type allocation should be investigated by fairness and characterized by solution of optimization problems. The asymptotic behaviour (e.g. fluid limits) of these models are also an inportant part of the research, because based on numerical investigations it seems the decay of performance parameters towards the asymptotic limit are well characterizable and useful for developing simple dimensioning rules.

Open problems and research goals:

  • Creating efficient analytical and/or numerical methods for solving the discriminative processor sharing with access rate limitations
  • Characterizing the decay of the scaled performance metrics (e.g. scaled average service times) between the two extreme cases, the access rate limitless case and the fluid limit case.
  • Finding dimensioning rules for buffering space and/or transmision rates based nt he performance evaluation results.

Előírt nyelvtudás:

  • English

További elvárások:

  • strong mathematical background, especially in stochastic modelling and probability theory
  • experience in computer programming

Felvehető hallgatók száma: 1