Model And Solution Algorithms For Multi-objective Closed-loop Logistics Network

Recently, due to the increasing environmental and social concerns along witheconomic benefits, an increasing number of companies focused on reverse logistics inaddition to forward logistics. Forward logistics encompasses material supply,production, distribution and consumption. Reverse logistics encompasses the flow ofused products includes collection, inspection/separation, recovering, disposal, andredistribution. Such a network combination is considered as a closed-loop logistics(CLL) network.In spite of the growing body of literature on CLL, little effort has been devoted tosynthesizing the researches on CLL. A literature review is a valid approach and anecessary step towards structuring of a research field, and forms an integral part of anyresearch conducted and also helps to identify the conceptual content of the field and cancontribute to theory development. In this research work, the content analysis method isused to review CLL literature. A content analysis can capture formal aspects as well ascontent aspects by applying a systematic procedure. To the best of our knowledge, thisis the only research that presents a literature review on closed-loop logistics based onthe modeling and solution approaches.At a planning stage, different decision making levels arise in the CLL networkdepending on the time horizon, namely, strategic, tactical, and operational. This researchfocuses on the facility location allocation problem which belongs to the strategic level.This type of problem includes designing the logistics configuration, determining thenumber and capacity of facilities, selecting the location of facilities, assigning facilitiesand determining the flow of quantity among facilities and consumers.Unlike previous studies which consider multi-product or multi-period inmulti-objective function problems; this research work formulates a multi-productmulti-period multi-objective mixed integer nonlinear programming with regard tofacility expansion. This multi-stage CLL network model includes a plant, a distributioncenter, retailers, a collection center, a recovery center, and a recycling center.A novel interactive fuzzy goal programming (IFGP) is proposed to solve the abovemodel for handling multiple conflicting objectives and the imprecise nature of decision makers’ aspiration levels for the goals. The solution approach controls the searchdirection via updating the aspiration level and the membership function value of eachobjective. The computational results show-when compared with the two other existingmethods in the literature-that the proposed solution approach can make the highersatisfactory degree for decision makers by providing more robust, reliable, and efficientsolutions.In the next step, a priority-based genetic algorithm is presented by using a straightchromosome and its encoding method. Finally, to assess the solution method of straightchromosome, the result is compared to the initial chromosome using different testscenarios. The numerical result shows that the straight chromosome outperforms theinitial chromosome at least in terms of the quality of the final solution.It is very common in the CLL that some of the data cannot be absolutely reliable.This research work addresses the uncertainty parameters in closed-loop logisticsnetwork using three well known robust counterpart optimization formulations. Finally,this research work assesses the results of three formulations with each other usingdifferent test scenarios. Bertsimas’ formulation is more suitable for an uncertainclosed-loop logistics network. Firstly, it does not increase the size of the modelconsiderably and it preserves its linearity. Secondly, the numerical result shows that itoutperforms the other two formulations in terms of quality of the final solution, CPUtime, and the level of conservatism along with the feasibility for the robust optimizationformulation.The main contributions of this research work can be described as follows:Develop a multi-objective multi-period multi-product closed-loop logistics networkwith regard to facility expansion.Propose a novel interactive fuzzy goal programming for handling multipleconflicting objectives and the imprecise nature of decision makers’ aspiration levelsfor the goals.Solve the model with a new priority-based genetic algorithm using a straightchromosome and its encoding method.Address the uncertainty parameters in the mathematical model using three wellknown robust counterpart optimization formulations.