A Comparative Study on Fashion Demand Forecasting Models with Multiple Sources of Uncertainty

Shuyun Ren, Hau Ling Chan, Pratibha Ram

Research output: Contribution to journalArticlepeer-review


Fast fashion is a timely, influential and well observed business strategy in the fashion retail industry. An effective fast fashion supply chain relies on quick and competent forecasts of highly volatile demand that involves multiple stock keeping units. However, there are multiple sources of uncertainty, such as market situation and rapid changes of the fashion trends, which makes demand forecasting more challenging. Therefore, it is crucial for the fast fashion companies to carefully select the right forecasting models to thrive and to succeed in this ever changing business environment. In this study, we first review a selected set of computational models which can be applied for fast fashion demand forecasting. We then perform a real sale data based computation analysis and discuss the strengths and weaknesses of these versatile models. Finally, we conduct a survey to learn about the perceived importance of different demand forecasting systems’ features from the fashion industry. Finally, we rank the fast fashion demand forecasting systems using the AHP analysis and supplement with important insights on the preferences on the demand forecasting systems of different groups of fashion industry experts and supply chain practitioners.
Original languageEnglish
Pages (from-to)335-355
Number of pages21
JournalAnnals of Operations Research
Issue number1-2
Publication statusPublished - 23 Apr 2016


  • Industrial applications
  • Uncertainty demand forecasting systems
  • Computational models
  • AHP analysis
  • fast fashion
  • RFID


Dive into the research topics of 'A Comparative Study on Fashion Demand Forecasting Models with Multiple Sources of Uncertainty'. Together they form a unique fingerprint.

Cite this