Possible research topics at Honours, Master or PhD level, in the area of:

A. Automated and distributed quantitative stochastic simulation

The research is focused on designing new methodologies and tools for performance modeling and evaluation of Future Internet and other Next Generation Networks by means of automated quantitative stochastic simulation. One of goals is to extend the functionality of Akaroa2, a unique controller of quantitative stochastic simulation, designed by our Simulation Research Group. Examples of some possible research topics in this area are listed below.

1. A1. Distributed stochastic simulation on PlanetLab.
2. A2. Comparative studies of existing sequential methods of analysis of steady-state means and their extensions for MRIP (Multiple Replications in Parallel) simulation
3. A3. Beyond mean-value analysis of output data (e.g. variance, quantiles) during sequential stochastic simulation with MRIP.
4. A4. Quantitative simulation of processes with heavy-tailed probability distributions and fractal processes
5. A5. Investigations of transient effects in steady-state simulation of networks
6. A6. New measures of coverage of final results from sequential simulation
7. A7. Study of different stopping criteria for sequential analysis of simulation output data (delta, delta/X, MSE, variance, etc) in MRIP, with the emphasis on heavy-tailed and self-similar processes.
8. A8. Automated sequential estimation of rare events during quantitative simulation, including MRIP simulation
9. A9. The initial transient phase in multi-parameter analysis of simulation output data, including the case of MRIP.
10. A10. Speeding up simulation by applying MRIP as the Variance Reduction Technique (VRT)
11. A11. Indirect Estimation as a VRT dependent on the stopping criterion of sequential simulation.
12. A12. Generation of pseudo-random correlated sequences by e.g. TES and ARTA
13. A13. Experimental studies of poor coverage caused by too short simulation runs and rules of thumb for improving coverage of the final results.
14. A14. Simulation of networks with fractal teletraffic: methods and their properties, including MRIP simulation.
15. A15.Comparative studies of new trends in pseudo-random number generation: single instruction generators, non-lattice congruential generators, etc.

B. Future Internet and Next Generation Networks, their design, performance modeling and evaluation by means of theory, simulation and experimentation

Research can involve studying analytical/mathematical models and simulation models linked with Akaroa2, with possible applications of global experimental networking infrastructures - accessible via PlanetLab NZ

Last time updated: April 6, 2013.