Mark Squillante, IBM

Abstract

Motivated by a wide variety of applications arising in practice, such as computer capacity planning and business process management, we develop a mathematical framework for determining the optimal resource capacity of each station composing a stochastic network.  The problem is mathematically intractable in general and therefore previous work typically resorts to either simplistic analytical approximations or time-consuming simulation-based optimization methods.  Our solution framework includes an iterative methodology that relies only on the capability of observing the queue lengths at all network stations for a given resource capacity allocation.  We theoretically investigate this proposed methodology for single-class Brownian tree networks and illustrate the use of our framework and the quality of its results through computational experiments.

Joint work with A. B. Dieker and S. Ghosh.

See: https://theory.cs.rutgers.edu/theory_seminar