A framework for early warning and proactive control systems in food supply chain networks Y. Li a , M.R. Kramer a, *, A.J.M. Beulens a , J.G.A.J. van der Vorst b a Information Technology Group, Wageningen University, Hollandseweg 1, 6706 KN Wageningen, The Netherlands b Operations Research and Logistics Group, Wageningen University, Hollandseweg 1, 6706 KN Wageningen, The Netherlands 1. Introduction A Dutch food company encountered a problem in their chicken supply chain; too many chickens died during transport to the slaughter house. This problem is known as Death-On-Arrival (DOA). The company continuously incurred extra costs because those dead chickens could not be used in the slaughter process and had to be disposed. Managers in this supply chain were not sure about the cause of this problem. However, the monitoring system in this supply chain recorded data associated with various factors (operational, environmental, etc.) and properties of chickens at various stages of the supply chain. So why not take advantage of those recorded data to try to predict and prevent DOA? The DOA problem is an example of a type of problems encountered in Food Supply Chain Networks (FSCN). In our research, we aim at building Early Warning and Proactive Control (EW&PC) systems to tackle such problems. Such systems are knowledge-based, data- and model-driven decision support systems (DSS) that are designed to assist managers in prediction and mitigation of problems associated with food products in FSCN. With these systems, managers can find causes for problems related to performance indicators of food products. The system analyses existing data, information, and knowledge available in FSCN and uses Data Mining (DM) to generate decision support models for the prediction of potential problems. Managers can in turn use these models in combination with their expertise in FSCN for decision- making. Due to characteristics of FSCN, EW&PC systems need a different architecture from traditional DSSs. In FSCN we have to deal with various types of problems, such as variability of quality and quantity of supply, shelf life constraints for raw materials, intermediates and finished product, variable process yield in quantity and quality due to biological variations, seasonality, random factors connected with weather, pests, other biological hazards [47]. In many cases, encountered performance problems are new to managers. The causes of such problems then need to be explored followed by the generation and evaluation of decision alternatives. Traditional architectures of DSSs (e.g. [29,36]), which are based on fixed models, are less suitable for EW&PC in FSCN. If encountered problems have not been included in those fixed models, managers can get little help. In this paper, we present a novel framework for EW&PC systems in FSCN. Such systems act as tools to solve a wide range of problems in FSCN based on available data. To our knowledge, such a framework has not been reported in literature before. We obtained our framework through several research steps (see Fig. 1). First, we formalized our ideas about EW&PC systems, and searched for relevant knowledge from literature on FSCN, DM, and DSS. Then we identified generic steps for realizing EW&PC with DM. In parallel, we conducted several case studies in the food industry, of which we use the DOA case to illustrate our findings. From case studies and DSS literature we derived various requirements on Computers in Industry 61 (2010) 852–862 ARTICLE INFO Keywords: Early warning Proactive control Data mining Knowledge management ABSTRACT It is inherent to food supply chain networks that performance deviations occur occasionally due to variations in product quality and quantity. To reduce losses, one wants to be informed about such deviations as soon as possible, preferably even before they occur. Then it is possible to take actions to prevent or reduce negative consequences. In practice, extensive operational data is recorded, forming a valuable source for early warning and proactive control systems, i.e. decision support systems for prediction and prevention of such performance problems. Data mining methods are ideal for analyzing such data sources and extracting useable information from them. In this paper, we present a novel framework for early warning and proactive control systems that combine expert knowledge and data mining methods to exploit recorded data. We discuss the implementation of a prototype system and the experiences from a case study regarding the applicability of the framework. ß 2010 Published by Elsevier B.V. * Corresponding author. E-mail address: Mark.Kramer@wur.nl (M.R. Kramer). Contents lists available at ScienceDirect Computers in Industry journal homepage: www.elsevier.com/locate/compind 0166-3615/$ – see front matter ß 2010 Published by Elsevier B.V. doi:10.1016/j.compind.2010.07.010