ZhuoDai Xiaoting Zheng
Schoolof Electronic Commerce, Jiu Jiang University,551 Qianjin East Road,332005, Jiu Jiang City, JiangXiProvince,
Correspondingauthor: Zhuo Dai
Abstract:The design of closed-loop supply chain network isone of the important issues in supply chain management. This research proposesa multi-period, multi-product, multi-echelon closed-loop supply chain networkdesign model under uncertainty. Because of its complexity, a solution frameworkwhich integrates Monte Carlo simulationembedded hybrid genetic algorithm, fuzzy programming and chance-constrainedprogramming jointly deal with the issue. A fuzzy programming and chance-constrainedprogramming approach take up the uncertainty issue. Monte Carlo simulationembedded hybrid genetic algorithm is employed to determine the configuration ofCLSC network. Parameters of GA are chosen to balance two aims. One aim is thatthe best value is global optimum, that is, maximum profit. The other aim is thatthe computational time is as short as possible. Non-parametric test confirmsthe advantage of hybrid GA. Then, the validity of Monte Carlo simulation embedded hybrid genetic algorithm is verified. The impactsof uncertainty in disposed rates, demands, and capacities on the overall profitof CLSC network are studied through sensitivity analysis. The proposed model is effective in designing CLSC network underuncertain environment.
Keywords: close-loop supply chain; network design;uncertainty; Monte Carlo simulation embedded hybridgenetic algorithm; fuzzy programming; chance-constrained programming
1. Introduction
An efficient supply chain can reduce costs and increase the profitof a company. One of the most important aspects of supply chain networkmanagement is supply chain network design. The problem of supply chain design whichconsiders both the efficiency and the risk was put forward by Huang and Goetschalckx(2014). They identified all Pareto-optimal configurations efficiently using branchand reduce algorithm.
In recentyears, people have begun to pay attention to environmental and socialresponsibilities and the economic benefits of used products, which make peopleattach importance to reverse supply chain network (Meade et al. 2007). Becauseof interdependent decisions in forward and reverse supply chain network,considering them separately lead to sub-optimal results (Pishvaee and Torabi2010). Therefore, decisions of forward and reverse supply chain network shouldbe considered simultaneously (Lee and Dong 2008). When a forward and reverse logisticsnetwork is considered simultaneously, a closed-loop supply chain (CLSC) networkwill be established. The aim of CLSC network is setting up an efficient systemfor bidirectional material flows considering environmental and economic effects.
Another important issue in designing supplychain network is uncertainty. The longer the time is,the higher the uncertainties are. In the CLSCnetwork, the uncertain problem is more serious due to the inherent uncertainty ofreverse logistics network, which is caused by uncertain factors such asquantity and quality of the used products (Pishvaee et al. 2011). Furthermore,uncertainties will be magnified through combinations and interactions among theabove uncertainties. Therefore, the issues of uncertainties should beconsidered and solved.
To address the issues of uncertainties, inexact optimization techniquesare employed in a mix-integer programming framework, for example, interval programming(Zhang et al. 2011), fuzzy programming (Vahdani et al. 2012) and stochastic programming(Kerachian and Karamouz 2007). Environmental coefficients are often fuzzy for mostproblems of supply chain network design because of incomplete information.Conventional methods cannot solve these problems. Zadeh (1965) put forwardfuzzy set theory. Fuzzy set theory has been applied to many fields such asproduction planning of supply chain, design of supply chain and so on. Liangand Cheng (2009) studied manufacturing and distribution planning decisionproblems in supply chains using fuzzy set theory. Peidro et al. (2007) proposeda fuzzy mixed integer linear programming model for supply chain planning. Xuand Zhai (2010) considered a supply chain coordination problem under fuzzyenvironmental constraints. They found that the expected profit of supply chainin a coordination situation is more than the profit in a non-coordinationsituation. Stochastic programming handles uncertain problems whose parameters’ probabilitydistributions are known (Liu and Sahinidis 1996). There are two types ofapproaches in stochastic programming. The first type of stochastic programmingapproach is recourse programs. The second type of stochastic programmingapproach is chance-constrained programming, which was proposed by Charnes andCooper (1959).
