Applied Probability and Simulation

ISyE Professor Sigrun Andradottir, ISyE Ph.D. student Mi Lim Lee, and Emory University BIostatistician Azhar Nizam standing in front of world map

ISyE Professor Sigrun Andradottir, ISyE Ph.D. student Mi Lim Lee, and Emory University BIostatistician Azhar Nizam collaborate on pandemic influenza simulation research.

ISyE’s research in applied probability and simulation ranges from theory to application. Current methodological research in applied probability includes:

  • Approximations of queuing networks.
  • Rare events (e.g., large deviations, extreme values, and tail asymptotics) in queuing systems, Gaussian processes, and heavy-tailed environments.
  • Duality, mixing times, and phase transitions in (controlled or noncontrolled) Markov processes.
  • Design and analysis of algorithms and related combinatorics.
  • Inventory theory.

On the simulation side, one finds projects that seek to improve the efficiency and validity of complicated simulations. Examples of such efforts include:

  • Optimization of simulation models (e.g., via adaptive random search, ranking and selection, or multinomial selection).
  • Simulation of stochastic processes with strong dependencies.
  • Sequential procedures for estimating quantiles.
  • Variance reduction methods to estimate option prices.
  • Sampling from model manifolds.

Much of the applied probability and simulation research is on modeling and analysis of real-world systems involving uncertainty. Recent projects include:

  • Allocation of resources such as clerks, laptops, and voting machines for elections.
  • Option pricing and dynamic portfolio choice.
  • Flexible server assignment in queuing networks.
  • Modeling and surveillance of infectious disease outbreaks.
  • Real-time management of complex resource allocation systems.
  • Routing and design of the overhead AGV system in a wafer fab.
  • Pricing in networks of queues.
  • Probabilistic modeling of fire ant (Solenopsis invicta) bivouac towers.
  • Admission control in loss networks.
  • Estimation of the probability of winning a random round-robin tournament.
  • Robust design of unit-load storage systems.
  • Multi-armed bandit problems. 

Georgia Tech Affiliations

Labs