الگوریتم اصلاحی با برچسب چند هدفه برای مدل سازی زنجیره تامین

Purpose—The purpose of this paper is to provide a multi-objective label correcting algorithm (MLC) to solve supply chain modeling problems. Design/methodology/approach—In the study, the proposed approach extends label correcting algorithm to apply to multi-objective problems where various conflicting objectives have to be considered simultaneously so as to make a trade-off between different criteria at the same time. The algorithm performs a forward search from a selected output point to all accessible points. It processes the stack of nodes based on the last-in-first-out (LIFO) rule. For the serial way, a node in the stack is pulled out for exploration and its subsequent node(s) is (are) pushed into the stack in each iteration. Then the MLC determines and updates the paths. The algorithm terminates when all nodes in the stack are pulled. Findings—The paper demonstrates that the MLC methodology can successfully resolve the supplier selection problem by taking into account the preference of the decision makers. Originality/value—many researches had proposed that many areas of the industry as, for example, telecommunications, transportation, aeronautics, chemistry, mechanical, and environment, deal with multi-objective, where various conflicting objectives have to be considered simultaneously. The method presented in the paper will help future studies in modeling of supply chain.

Mithun et al. (2008) say that design of the chain should be able to integrate the various elements of the chain and should strive for the optimization of the chain rather than the entities or group of entities. In recent year, supply chain management is a holistic and strategic approach to demand, operations, procurement, and logistics process management (Mithun et al., 2008). Over the years, most firms have focused their attention on the effectiveness and efficiency of separate business functions. As new ways of doing business, however, a growing number of firms have begun to realize the strategic importance of planning, controlling, and designing a supply chain as a whole. Modeling is crucial work in supply chain since this is done in an effort to help firms capture the synergy of inter-functional and inter-organizational integration and coordination across the supply chain and to subsequently make better supply chain decisions. Supply chain modeling can be characterized as a primary method- or algorithm-oriented approach towards SCM. Supply chain model is often represented as a resource network. The nodes in the network represent facilities, which are connected by links that represent direct transportation connections permitted by the company in managing its supply chain. Supply chain modeling has to configure this network and to program the flows within the configuration according to a specific objective function based on algorithms (Swaminathan and Tayur, 1999). Therefore, supply chain can be modeled as a configurable and flow-programmable resource network. The network employs a completely different and very selective view of what is going on in the supply chain. But as literature and practice prove, it is a quite powerful way of improving the chain (Kotzab and Otto, 2003). Supply chain modeling offers short-, medium- or long-term optimization potentials. Elements within the optimization scope may be plants, distribution centers, suppliers, customers, orders, products, or inventories. The standard problems for supply chain modeling are formulated in the following manner. A set of goals should be achieved by minimizing the costs of transfer and transformation. In partial solutions, particular goals are selected, such as securing a certain service level to minimize lead time and maximize capacity utilization, or to secure availability of resources (Kotzab and Otto, 2003). The standard solutions in supply chain modeling can be found in the establishment of certain algorithms, which identify the optimal solution for the problem. The label correcting algorithm has been proven to be extremely efficient through solving a certain supply chain modeling problem—the shortest path problem (Angelica and Giovanni, 2001 and Angelica and Giovanni, 2002). In previous researches, many areas of the industry, for example, telecommunications, transportation, aeronautics, chemistry, mechanical, and environment, deal with multi-objective, where various conflicting objectives have to be considered simultaneously (Figueira et al., 2010). For example, the relationship between production and maintenance objectives has been considered as a conflict in management decision. If a decision maker uses single objective algorithm to optimize the production objective or the maintenance objective separately, the conflict may result in an unsatisfied demand or machine breakdowns. It is because the production services and maintenance services do not respect the requirement of each other (Berrichi et al., 2010).Due to these reasons, this paper develops a multi-objective label correcting algorithm to achieve a trade-off but consider different criteria at the same time.

A growing number of firms have begun to realize the strategic importance of planning, controlling, and designing a supply chain as a whole. In an effort to help firms capture the spirit of integration and coordination across the supply chain and to subsequently make better supply chain decisions, synthesize of past supply chain modeling efforts and a novel approach to modeling supply chain are required. The label correcting algorithm is proven to be extremely efficient for the shortest path problem. This paper develops the multi-objective label correcting algorithm (MLC) to solve supply chain modeling problems. The proposed approach extends label correcting algorithm to apply to multi-objective problems where various conflicting objectives have to be considered simultaneously so as to make a trade-off between different criteria at the same time. To industry, the MLC methodology can successfully resolve the supplier selection problem by taking into account the preference of the decision makers. To avoid Pareto scenario, the sensitivity analysis on the weights of preferences is implemented. Through sensitivity analysis, industry can recognize which preference is required to be focused in the multi-objective problems. For example, the nature of the production in this case focuses on reliability most. As the weight of reliability is determined, the changes of cost and time on weight have no impact on the optimal solution.