TR2013-003

Fast Runcurve Optimization based on Markov Decision Process


    •  Fujii, S., Yoshimoto, K., Ueda, K., Takahashi, S., Nikovski, D., "Fast Runcurve Optimization based on Markov Decision Process", Symposium of the Society of Instrumentation and Control Engineers of Japan (SICE), January 2013.
      BibTeX TR2013-003 PDF
      • @inproceedings{Fujii2013jan,
      • author = {Fujii, S. and Yoshimoto, K. and Ueda, K. and Takahashi, S. and Nikovski, D.},
      • title = {Fast Runcurve Optimization based on Markov Decision Process},
      • booktitle = {Symposium of the Society of Instrumentation and Control Engineers of Japan (SICE)},
      • year = 2013,
      • month = jan,
      • url = {https://www.merl.com/publications/TR2013-003}
      • }
  • MERL Contact:
Abstract:

This paper proposes a fast algorithm to solve runcurve optimization problem. We model the problem with Marcov decision process in which cost function is defined by a weighted sum of energy consumption and running time, select appropriate weight values, and compute optimal runcurve with dynamic programming. We have confirmed that the proposed algorithm could optimize runcurve of 2,000m within around 1 sec by simulation results.

 

  • Related News & Events

    •  NEWS    SICE 2013: publication by Daniel N. Nikovski and others
      Date: January 18, 2013
      Where: Symposium of the Society of Instrumentation and Control Engineers of Japan (SICE)
      MERL Contact: Daniel N. Nikovski
      Research Area: Data Analytics
      Brief
      • The paper "Fast Runcurve Optimization based on Markov Decision Process" by Fujii, S., Yoshimoto, K., Ueda, K., Takahashi, S. and Nikovski, D. was presented at the Symposium of the Society of Instrumentation and Control Engineers of Japan (SICE).
    •