MultiCraft International Journal of Engineering, Science and Technology Vol. 5, No. 2, 2013, pp. 93-109 INTERNATIONAL JOURNAL OF ENGINEERING, SCIENCE AND TECHNOLOGY www.ijest-ng.com www.ajol.info/index.php/ijest © 2013 MultiCraft Limited. All rights reserved Design of supply networks with optimized safety stock levels Eleonora Bottani 1 *, Gino Ferretti 1 , Roberto Montanari 1 , Giuseppe Vignali 1 , Francesco Longo 2 , Agostino Bruzzone 3 1 Department of Industrial Engineering, University of Parma, viale G.P.Usberti 181/A, 43124 Parma, ITALY 2 Simulation Team, ITALY 3 DIPTEM University of Genoa, via Opera Pia 15, 16145 Genova, ITALY * Corresponding Author: e-mail: eleonora.bottani@unipr.it, Phone +39-0521-905872, Fax +39-0521-905705 Abstract In this paper, we address two main issues. First, we determine, through a simulation model, the optimal size and distribution of the safety stocks in a supply network. The “optimal” size of safety stocks results from the minimization of the total logistics cost of the supply network, as a function of the safety stock coefficient (k). In particular, we define the “optimal” size of the safety stocks as the k value which minimizes the total logistics cost of the network. Second, the optimized values of k are used to run the same simulation model under different operating conditions of the network, which are obtained by introducing demand stochasticity, demand seasonality and lead time stochasticity. More precisely, once the optimal size of safety stocks has been set, we carry out further simulations, according to the design of experiments (DOE) procedure, and perform statistical analyses of the resulting outputs, to provide some insights about the design of the supply network under optimal safety stock level. The study is supported by a discrete-event simulation model, developed ad hoc and reproducing 4 different configurations of a fast moving consumer goods (FMCG) network. To run the model, real data related to the FMCG context were used. Results of this study can be useful to supply chain managers, to identify the optimal service level the network should deliver to customers, as well as to understand the behavior of supply networks under optimal safety stock level. Keywords: safety stock; supply network design; simulation model; design of experiments; optimization. DOI: http://dx.doi.org/10.4314/ijest.v5i2.7S 1. Introduction Supply chain management (SCM) is the process of integrating suppliers, manufacturers, warehouses, and retailers in a supply chain, so that goods are produced and delivered in the right quantities and at the right time, while minimizing costs as well as satisfying customer’s requirements (Cooper et al., 1997). Managing the entire supply chain is a key success factor for any business, as non-integrated manufacturing processes, non-integrated distribution processes and poor relationships with suppliers and customers inevitably lead to company failure (Chang and Makatsoris, 2001). Efficiently and effectively managing the supply chain involves different interrelated topics, namely (i) defining the supply chain (or supply network) structure, (ii) identifying the supply chain business processes and (iii) identifying the business components (Lambert, 2001). The first topic, in particular, encompasses a set of decisions concerning, among others, the number of echelons required and the number of facilities per echelon, the reorder policy to be adopted by echelons, the service level to be delivered to customers, the assignment of each market region to one or more facilities, and the selection of suppliers for sub-assemblies, components and materials (Chopra and Meindl, 2004; Hammami et al., 2008). Moreover, different supply chain configurations react differently to the bullwhip effect, and they result in different levels of safety stocks required; examples of case studies in the retail sector in which the inventory levels along the supply chain (under different conditions, external constraints and supply chain configurations) are investigated can be found in Bruzzone and Longo (2010) and Longo and Mirabelli (2008).