Analysis of Damaged Containers Stacking Strategies

Abstract

A container strategy based on information about container damage is considered in the paper. The container terminals operators frequently use special sections in container yard, dedicated only for the damaged containers. It is suggested in the paper that this strategy leads to the minimization of additional movements in container selection operations. At the same time, the fact that container owners do not repair all the damaged containers is proved in the paper. This leads to the necessity to select containers from the specialized section with additional movements. This selection can be avoided if the specialized section consists of the containers that will be certainly repaired. One of the hypotheses of the research is that the effectiveness of this strategy can be improved if a statistical classifier is used for the prediction of container repair. This classifier can be based on the information about the price of container repair and container exploitation time. The linear classifier educated on the small statistical data is considered in the paper. It is proved that such model could effectively classify 40 % of containers that will be repaired. In order to analyze the effectiveness of suggested strategies the simulation model is used. There are three strategies, that are considered in the paper: the specialized stack is used only for the container classified with statistical model; the specialized stack is used for all damaged containers; containers are stacked to the stack minimum positions. It is proved that the first two strategies increase the number of additional movements in the nonspecialized stack. The results of analysis have showed that the hypothesis of statistical classifier utilization in order to improve stacking strategy effectiveness is not practically useful. The most useful strategy in the case of the research is containers even distribution in the stack.