A multiobjective optimization-based neural network model for short-term replenishment forecasting in fashion industry

Wei Du, Sunney Yung Sun Leung, Chun Kit Kwong

Research output: Journal article publicationJournal articleAcademic researchpeer-review

20 Citations (Scopus)


A multiobjective optimization-based neural network (MOONN) model is proposed to tackle the short-term replenishment forecasting problem in fashion industry. Our approach utilizes a new multiobjective evolutionary algorithm called nondominated sorting adaptive differential evolution algorithm (NSJADE) to optimize the weights of neural networks (NNs) for the short-term replenishment forecasting problem, acquiring the forecasting accuracy while alleviating the overfitting effect at the same time. The presented NSJADE also selects the appropriate number of hidden nodes for the NN according to different short-term replenishment forecasting problems. Extensive experiments based on real fashion industry data are performed to validate the effectiveness of the developed model. Experimental results reveal that the performance of the proposed model is superior than several popular models for the short-term replenishment forecasting problem.
Original languageEnglish
Pages (from-to)342-353
Number of pages12
Issue numberP1
Publication statusPublished - 1 Jan 2015


  • Fashion sales forecasting
  • Multiobjective optimization
  • Neural network
  • Short-term replenishment forecasting

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Science Applications
  • Cognitive Neuroscience

Cite this