Development of bi-objective optimization model for supply chain network design using data envelopment analysis

In supply chain, there are several facilities including suppliers, manufacturing, warehouses, and distributors and vendors which develop the product, procure material, and move products, produce products, finally distribute finished products between sites. Understanding different aspects of facility location in supply chain network, such as operations, decision-making policies and relating them to performance measurements have been increasingly investigated in the last decade. The number of facilities, location, and capacity of the facility affects the performance of supply chain. Therefore facility location decision in the supply chain can be performed simultaneously with data envelopment analysis (DEA) efficiency measurement. Recently data envelopment analysis method has been used to measure the performance of decision-making units (DMUs), though, sometimes there are DMUs which their behavior are like a network that a classical DEA method cannot deal with, as in the case of supply chain network. To design the optimal supply chain network several objectives, such as cost, environment, social, coverage need to be considered. Facility efficiency is recently focused by some scholars as a new objective in supply chain network. With this, network DEA has been introduced as the best model to solve this problem. In this thesis, there are two objectives which should be achieved. First objective is obtaining an optimal DEA efficiency score simultaneously with facility location pattern for two stages supply chain are studied as a bi-objective model, and a trade-off between facility location cost with facility efficiency score alongside sensitive analysis in the supply chain are shown. Moreover, in the second objective Benders decomposition algorithm has been introduced as an effective approach for large-scale size problem. Several examples have been applied to verify and validate the effectiveness of proposed model and Benders decomposition algorithm. An example from the real case was considered to verify and validate the proposed model. One numerical example has been illustrated the effectiveness of Benders decomposition algorithm for the complicated problem. One example with standard data from Malaysian business has shown to depict the effectiveness of proposed approach for facility location-allocation problem as a mixed integer optimization problem. Furthermore, another example from the real case has compared the effectiveness of the Benders decomposition for the proposed model with a solution has been found from the original problem and solved with the CPLEX solver. In this regard, other simulation data for the large-scale cases in the range of real case also compared. Analysis of the results expressed acceptable performance of the developed model and proposed solution for different cases in different sizes. The developed model and solution method show excellence performance in terms of CPU time for the large scale. And in the last part conclusion and future works are presented.

Preview Text FK 2017 81 - IR.pdf Download (1MB) | Preview

Actions (login required)

View Item