A Hybrid Grey & ANFIS Approach to Bullwhip Effect in Supply Chain Networks HAKAN TOZAN OZALP VAYVAY Department of Industrial Engineering Turkish Naval Academy 34942 Tuzla / Istanbul TURKIYE htozan@dho.edu.tr Abstract: - Demand forecasting and decision making processes are among the key activities which directly affect the performance of successful supply chain networks. The variability of the demand information between the stages of the supply chains and the increase in this variability as the demand data moves upstream from the customer to consequent stages of the supply chain networks is called Bullwhip Effect. As demand pattern varies due to the field of activity and architecture of supply chains networks, determining the appropriate forecasting and order decision model for system interested in is complicated. This paper analyzes the response of bullwhip effect to a hybrid grey GM (1, 1) forecasting and ANFIS based order decision model under demand with relatively medium variation in a two stage supply chain network simulation. Key-Words: - Forecasting, Supply chain network, Bullwhip effect, ANFIS, Grey GM (1, 1). 1 Introduction Supply chain networks (ScNs) are multi stage complex dynamical systems consist of various involved organizations performing different processes and activities in each and consequent stages which are connected through upstream and downstream linkages to produce value in the form of products and services [1, 2]. Fig.1. A simple multistage ScN Bethinking of the definition exposes that the performance of a successful ScN system directly depends on accurate and appropriate demand information, as this vital data influences all decision making processes of ScN. The information flow in ScN consists of cumulative data about costs parameters, production activities, inventory systems and levels, logistic activities and many other related complex processes. But basically; in addition to the architecture, system performance of successful ScNs directly depends on accurate, constant, on time and appropriate demand information flow through the stages of the system grounding on the decision making process and estimated values obtained from the selected forecasting activities and decision making processes performed in each stage. The variability of the demand information between the stages of ScN and the increase in this variability as the demand data moves upstream from the customer to the consequent stages is called Bullwhip Effect (BE). Fig.2. Bullwhip effect in a two stage ScN This phenomenon triggers several system defects which directly influence total performance of ScNs. such as WSEAS TRANSACTIONS on SYSTEMS Hakan Tozan, Ozalp Vayvay ISSN: 1109-2777 461 Issue 4, Volume 8, April 2009