Application of Bayesian Networks to Architectural Optimisation

  • Authors:

  • Artem Parakhine;Tim O'Neill;John Leaney

  • Affiliations:

  • University of Technology, Sydney, Australia;University of Technology, Sydney, Australia;University of Technology, Sydney, Australia

  • Venue:
  • ECBS '07 Proceedings of the 14th Annual IEEE International Conference and Workshops on the Engineering of Computer-Based Systems
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

The field of optimisation covers a great multitude of principles, methods and frameworks aimed at maximisation of an objective under constraints. However, the classical optimisation can not be easily applied in the context of computer-based systems architecture as there is not enough knowledge concerning the dependencies between non-functional qualities of the system. Out approach is based on the simulation optimisation methodology where the system simulation is first created to assess the current state of the design with respect to the objectives. The results of the simulation are used to construct a Bayesian Belief Network which effectively becomes a base for an objective function and serves as the main source of the decision support pertaining to the guidance of the optimisation process. The potential effects of each proposed change or combination of changes is then examined by updating and re-evaluating the system simulation.