6th International Conference on Performance Evaluation Methodologies and Tools

Research Article

Online Optimization of Product-Form Networks

Download508 downloads
  • @INPROCEEDINGS{10.4108/valuetools.2012.250261,
        author={Jaron Sanders and Sem Borst and Johan van Leeuwaarden},
        title={Online Optimization of Product-Form Networks},
        proceedings={6th International Conference on Performance Evaluation Methodologies and Tools},
        publisher={IEEE},
        proceedings_a={VALUETOOLS},
        year={2012},
        month={11},
        keywords={gradient algorithm markov processes mixing times online performance optimization product-form networks stochastic approximation dynamic control},
        doi={10.4108/valuetools.2012.250261}
    }
    
  • Jaron Sanders
    Sem Borst
    Johan van Leeuwaarden
    Year: 2012
    Online Optimization of Product-Form Networks
    VALUETOOLS
    ICST
    DOI: 10.4108/valuetools.2012.250261
Jaron Sanders1,*, Sem Borst1, Johan van Leeuwaarden1
  • 1: Eindhoven University of Technology
*Contact email: jaron.sanders@tue.nl

Abstract

We develop an online gradient algorithm for optimizing the performance of product-form networks through online adjustment of control parameters. The use of standard algorithms for finding optimal parameter settings is hampered by the prohibitive computational burden of calculating the gradient in terms of the stationary probabilities. The proposed approach instead relies on measuring empirical frequencies of the various states through simulation or online operation so as to obtain estimates for the gradient. Besides the reduction in computational effort, a further benefit of the online operation lies in the natural adaptation to slow variations in ambient parameters as commonly occurring in dynamic environments. On the downside, the measurements result in inherently noisy and biased estimates. We exploit mixing time results in order to overcome the impact of the bias and establish sufficient conditions for convergence to a globally optimal solution.