컨테이너 선박 자동 선적계획을 위한 심층 학습 모형

Abstract

The stowage planning in containership becomes crucial to optimize the shipping logistics supply chain but complexity of it becomes very high especially for large containership like over 15,000 TEU. The quality of stowage varies from the level of skill of stowage planners, and it impacts a lot to the profit and loss. To overcome the problem, there were many studies in this field for the stowage automation but none of them becomes commercialized due to the NP-Complete problem. This study defines the process of stowage planning automation and its evaluation criteria and builds two different deep learning models - one is the Block Stowage and another is Lashing Force prediction. The Reinforcement Learning is used for Block Stowage to allocate stowage spaces for multiple ports and the Supervised Learning is studied for Lashing Force prediction which is one of the most important aspects in terms of cargo safety as well as utilization, especially for large containership such as more than 10,000 TEU. It is a challenge for stowage planners when large containerships are in the last port of region because mostly the ship is full and the stacks on deck are very high. This study also conducts training and testing the models and provides the results in comparison with stowage planning tool.