Forecasting hierarchical time series in supply chains: an empirical investigation

Abstract

Demand forecasting is a fundamental component of efficient supply chain management. An accurate demand forecast is required at several different levels of a supply chain network to support the planning and decision-making process in various departments. In this paper, we investigate the performance of bottom-up, top-down and optimal combination forecasting approaches in a supply chain. We first evaluate their forecast performance by means of a simulation study and an empirical investigation in a multi-echelon distribution network from a major European brewery company. For the latter, the grouped time series forecasting structure is designed to support managers’ decisions in manufacturing, marketing, finance and logistics. Then, we examine the forecast accuracy of combining forecasts of these approaches. Results reveal that forecast combinations produce forecasts that are more accurate and less biased than individual approaches. Moreover, we develop a model to analyse the association between time series characteristics and the effectiveness of each approach. Results provide insights into the interaction among time series characteristics and the performance of these approaches at the bottom level of the hierarchy. Valuable insights are offered to practitioners and the paper closes with final remarks and agenda for further research in this area.

KEYWORDS:

• Supply chain forecasting
• forecast combination
• hierarchical forecasting
• grouped time series forecasting
• time series characteristics

Acknowledgements

We would like to thank all our colleagues, friends and family who contributed to the development of this paper by providing us with feedback during the several research phases. Also, we thank the anonymous referees for their constructive comments which greatly increase the quality of our research.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 GF can be considered as a special case of HF. Depending on the demand structure of the SC, HF or GF methodology might be used.

2 Additional forecasting error metrics (RMSE, MAPE, MAE, MASE and ME), are also available from the corresponding author on request and through Shiny platform: https://supplychainanalytics.shinyapps.io/empirical_beverage_study/.

3 We restrict extrapolation of the findings regarding the association of time series characteristics on the forecasting performance, only on those time series which have similar or closely related summary statistics to the data provided in Table 3 and Figure B1.

Additional information

Funding

We share our gratitude to the Serbian government since this research has been supported by the Ministry of Education, Science and Technological Development through [project no. 451-03-68/2020-14/200156]: ‘Innovative scientific and artistic research from the FTS (activity) domain’.

Notes on contributors

Dejan Mircetic

Dejan Mircetic is a Teaching Assistant at the University of Novi Sad/Faculty of Technical Sciences (Serbia), where he teaches several subjects related to supply chains and logistics. He holds an MSc in logistics management and a PhD in transport engineering. His PhD research deals with supply chain forecasting with a focus on hierarchical forecasting. Dejan research interests primarily relate to demand forecasting and inventory management with a special emphasis on the development of quantitative models. He is a member of various scientific and commercial projects founded by different programs (Interreg, USAID, COST, Visegrad, CEEPUS, and Erasmus plus). Besides, he has been rewarded with several mobility grants in many European Universities and countries including: Italy, Slovenia, Slovakia, Poland, the Czech Republic, Hungary, and Croatia. For his scientific research, he has received several rewards, among them the prestigious Sakura reward from his excellency Japanese ambassador in Serbia Toshio Cunozaki and the Japan Tobacco International company.

Bahman Rostami-Tabar

Bahman Rostami-Tabar is a senior lecturer of Management Science at Cardiff Business School (CARBS), Cardiff University, U.K. He holds a Ph.D. in Industrial Engineering from the University of Bordeaux, France. His research involves developing innovative methodologies in forecasting for social good and analysing the value of forecasting in decision making. Bahman launched the ‘Democratising Forecasting’ in 2018 sponsored by the International Institute of Forecasters (IIF) to deliver a series of workshops in developing economies which promotes the importance of forecasting and ‘train the trainers’ on forecasting principles using R. This initiative has recently been extended to train over 100 data analysts within the National Health Service (NHS, UK) in collaboration with the NHS-R community. Bahman has also created the Forecasting for Social Good (FSG) initiative sponsored by IIF that aims to explore and expose how (where) forecasting can have some positive societal impact. The objective is of this initiative threefold: (i) expose the importance of forecasting for social good, (ii) capture the otherwise dispersed current state of knowledge and set the agenda to drive developments (iii) create and establish research collaborations and partnership at the national and international level.

Svetlana Nikolicic

Svetlana Nikolicic is an Associate Professor at the Faculty of Technical Sciences – Department for Traffic and Transportation, University of Novi Sad, Serbia. Her Field of Academic Expertise is logistics and intermodal transportation. Her research interests are: the logistics organisation in company, modelling and improvement logistics processes and logistics performances in the manufacturing and trading companies. In recent years, she has been studying the impact of IT on logistics performance in retail supply chains. She was a member of the research teams in over 10 projects, related to logistics topics, most of which have been funded by the Serbian Ministry of Science. She has contributed about 10 articles to academic and professional journals and presented over 30 conference presentations on logistics topics in the country and abroad.

Marinko Maslaric

Marinko Maslaric is Associate Professor at the Novi Sad University (Serbia). He graduated as MSc in logistics management and defended his Ph.D. thesis in transport engineering. He is employed at the Faculty of Technical Sciences since 2005 where he teaches courses on Intermodal Transport and Supply Chain Management at the Department for Transport and Traffic Engineering. His main research fields are: supply chain management, logistics, transport engineering, optimisation and simulation in logistics. He has been granted by various EU programmes for scholarships to perform mobility in a number of EU universities from Slovenia, Italy, Finland, Slovakia, Poland, Czech Republic, The Netherlands, Hungary, Croatia, Romania, Bulgaria. He was included in several national scientific projects financed by the Serbian Ministry of Education and Science, as well as several international projects funded by H2020, Interreg, COST, Visegrad, and CEEPUS program. He is the author of more than 80 scientific papers.