Modeling And Optimization Of Container Drayage Transportation Problem With Objective Of Low Carbons

A report by the UN Intergovernmental Panel on climate change said:over the past 50 years, greenhouse gases emissions into the atmosphere were the main reason of the rising of the average global temperatures, and global warming will cause sea levels rising, frequency of extreme weather and other adverse natural phenomena, thus affecting the human production, life and even threaten the survival of mankind further. Logistics and transport industry is part of the main sources of greenhouse gases such as CO2. While the CO2 emissions of road freight are the important proportion of the total CO2 emissions in the logistics transportation. At present, container is the main form of freight transportation. Therefore, container transportation by road is an important aspect of road freight transport. Thus, research on container transportation by road with objective of low carbon will be friendly to the environment.Usually, at a long distance level, containers are mainly transported by vessels or by rains. While at a short distance level, container transportation is performed by trucks. We call the segment of transportation by trucks between customers and terminals as drayage. Container drayage transportation can provide door-to-door services. Although the transportation distances by trucks are short, the total costs per TEU (Twentyfoot Equivalent Unit) per kilometer are relatively high. Moreover, this transportation mode is the key source of road congestion, shipment delays, and disruptions. Therefore, it is very important to improve the efficiency of container transportation.This thesis studies a container drayage problem. We not only consider the binary time windows attributions of the transportation task and empty containers reposition, but also consider the effects of the container drayage transportation on the environment. We use the total carbon emissions of trucks in container drayage transportation instead of the traditional objective function that takes the total travelling time of trucks or the drayage operation costs as the objective function, and then we also analysis the container drayage transportation problem with objective of low carbons. The researches of this thesis are as follows:(1) Container drayage transportation problem with given speeds of trucks is studied. Based on a determined-activities-on-vertex (DAOV) graph, a mixed-integer linear programming model of the drayage transportation problem is built. It is tested that the calculating speed and precision of solving this model by commercial optimization software LINGO. This model is further compared with traditional models which minimize the total working time. The effects of the width of time windows, the ratio of inbound containers and the ratio of empty containers are analyzed. The results indicate that this method could provide high-precision solutions of realistic-sized instances quickly. It could also decrease carbon emissions significantly and hence has important academic meanings.(2) Container drayage transportation problem with variable speeds of trucks is studied. A determined-activities-on-vertex (DAOV) graph is extended. Then, based on the extended DAOV graph, this problem is formulated as a mixed integer nonlinear programming model which takes speeds of trucks as variables. A speed discretization based method is designed and the model is transferred into a mixed integer linear programming model. Results of experiments indicate that this method could provide near optimum solutions of realistic sized instances in a short period. The gap between such near optimum solutions and the theoretical optimum solutions of medium and small sized instances which optimum solutions could be provided by other methods is very small (usually less than 0.1 percent). The comparison with the problems in which speeds are given further validates the model and method.(3) Container drayage transportation problem with variable speeds of trucks is studied. Based on the extended DAOV graph in the last section, this problem is formulated as a mixed integer nonlinear programming model which takes travelling time of trucks as variables. A time windows discretization based method is designed and the model is transferred into a pure integer (0-1) linear programming model. Results of experiments indicate that this method could provide near optimum solutions of realistic sized instances in a short period. The gap between such near optimum solutions and the theoretical optimum solutions of medium and small sized instances which optimum solutions could be provided by other methods is very small (usually less than 0.1 percent). The comparison with speed discretization based method validates the time windows discretization based method.