Mathematical models and meta heuristics for capacitated green vehicle routing problems development and analysis

Abstract

The concepts of green economics and sustainability practices have been newlinereligiously shaping the futuristic supply chain network designs. The newly refined newlinegreen policies demand a changed way of conducting business and are rewriting newline newlinethe traditional profit centric economy into a more responsible and socio- newlineenvironmental centric one. In the supply chain design context, the transportation newline newlinesector seeks an obvious attention as a consequence of its overwhelming newlinedependency on fossil fuels. The green vehicle routing problem enjoined by the newlineglobal sustainability conscience is a seminal work on the contemporary green newlineresearch in supply chain management. Inspired by the replete pollution-routing newlineliterature, the present research acknowledges the significance of redesigning the newlinelogistic networks in a low emission perspective. newlineThis research proposes three logistic models to address the capacitated newlinevehicle routing problems namely, (i) capacitated vehicle routing problem - newlinepickup/collection only, (ii) multi-depot vehicle routing problem delivery only, newlineand (iii) location-routing problem with simultaneous pickup and delivery. The newlinefirst model generates classical, distance centric and economically viable route newlineplans. In the second model, the environmental friendly routes are generated. The newlinethird model is slated to bring a trade-off between the first two models. The three newlinemodels are formulated as integer linear programming models and are solved with newlinebranch-and-bound based exact algorithm. Owing to the computational time newlinecomplexity of the NP-hard problem for generating time bound results, an Ant newlineColony Optimisation (ACO) algorithm is designed to solve the models. The newlineperformance of the algorithm is further enhanced by introducing a powerful newlineVariable Neighbourhood Search (VNS). The first phase of the research addresses a milk collection problem from newlinethe Indian sub-continent. The work is motivated by the opportunity to redesign an newlineexisting logistic plan in a dairy industry.