Low-carbon multi-objective location-routing in supply chain network design
CHALLENGE - Traditional supply chain (SC) modelling tends to focus on singular objectives, with a predominant focus on cost and within this discipline location-routing problems (LRP) are one of the most researched categories in recent years. This PhD thesis tackles cost and environmental impact in the form of carbon emissions, focusing on low-cost low-carbon structure for the demand side of supply chain networks. The case SC network has 2 processing plants, 6 Distribution Centers and 22 retailers.
SOLUTION - The solution approach to the three-layer MO-LRP is DoE-guided. DoE generates the initial population for the optimisers in modeFRONTIER. The 20 selected results obtained from MOGA-II, MOPSO and NSGA-II optimisers are separately examined with regard to their Pareto efficiency and the results are presented. The selected Pareto efficient realistic results are evaluated using TOPSIS. In order to present the best routing patterns, a geographical representation of selected results are shown on the map of Ireland.
BENEFITS - Two and three layers of a SC network are considered and two generic multi-objective low-carbon location-routing models are designed and developed with the use of modeFRONTIER. The three-layer MO-LRP results offer optimal facility location and vehicle-routing decision. It is found that modeFRONTIER's NSGA-II is the best optimizer as compared with the other optimizers.