Computational Management Science

, Volume 2, Issue 4, pp 309–338 | Cite as

An application of the neural network energy function to machine sequencing

  • Komgrit Leksakul
  • Anulark Techanitisawad
Original Paper

Abstract.

We apply a neural network approach for solving a one-machine sequencing problem to minimize either single- or multi-objectives, namely the total tardiness, total flowtime, maximimum tardiness, maximum flowtime, and number of tardy jobs. We formulate correspondingly nonlinear integer models, for each of which we derive a quadratic energy function, a neural network, and a system of differential equations. Simulation results based on solving the nonlinear differential equations demonstrate that our approach can effectively solve the sequencing problems to optimality in most cases and near optimality in a few cases. The neural network approach can also be implemented on a parallel computing network, resulting in significant runtime savings over the optimization approach.

Keywords:

Media planning Advertising management Intervention analysis  Transfer function models 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer-Verlag Berlin/Heidelberg 2005

Authors and Affiliations

  • Komgrit Leksakul
    • 1
  • Anulark Techanitisawad
    • 1
  1. 1.Industrial Systems Engineering ProgramSchool of Advanced Technologies, Asian Institute of TechnologyKlong LuangThailand

Personalised recommendations