“交通·未来”第18期:基于机器学习的港口国监督检查选船和检查员排班预测优化模型
Posted 当交通遇上机器学习
tags:
篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了“交通·未来”第18期:基于机器学习的港口国监督检查选船和检查员排班预测优化模型相关的知识,希望对你有一定的参考价值。
基于机器学习的港口国监督检查选船和检查员排班预测优化模型
Efficientinspection of ships at ports to ensure their compliance with safety andenvironmental regulations is of vital significance to maritime transportation.Given that maritime authorities often have limited inspection resources, weproposed three two-step approaches that match the inspection resources with theships, aimed at identifying the most deficiencies (non-compliances withregulations) of the ships. The first approach predicts the number ofdeficiencies in each deficiency category for each ship and then develops aninteger optimization model that assigns the inspectors to the ships to beinspected. The second approach predicts the number of deficiencies eachinspector can identify for each ship and then applies an integer optimizationmodel to assign the inspectors to the ships to be inspected. The third approachis a semi-“smart predict then optimize” (semi-SPO) method. It also predicts thenumber of deficiencies each inspector can identify for each ship and uses thesame integer optimization model as the second approach, however, instead ofminimizing the mean squared error as in the second approach, it adopts a lossfunction motivated by the structure of the optimization problem in the secondapproach. Numerical experiments show that the proposed approaches improve thecurrent inspection efficiency by over 4% regarding the total number of detecteddeficiencies. Through comprehensive sensitivity analysis, several managerialinsights are generated and the robustness of the proposed approaches is validated.
鄢然:香港理工大学2018级研究型硕士生,2020级博士生。研究方向为航运管理与政策、航运数据分析、绿色航运管理。她的研究(第一作者)已被Transportation Research Part B,Journal of Computational Science, Transportation Research Part E, MaritimePolicy & Management, and Asia-Pacific Journal of Operational Research等期刊发表/接收,另有4篇论文(第一作者)在审。同时,她担任包括Transportation Research Part E,Ocean Engineering, Maritime Policy & Management和Journal of Air TransportManagement等多个SCI/SSCI期刊的审稿人。
讲座时间:2021年1月29日晚上19:00
讲座地点:腾讯会议ID:898 520 015
B站同步直播链接:https://live.bilibili.com/22297847(扫描下方二维码进入直播间)
第19期分享主题:待定
时间:待定
主讲人:待定
一个人走得快,但一群人走得远!期待我们共同成长为交通领域未来之星!
以上是关于“交通·未来”第18期:基于机器学习的港口国监督检查选船和检查员排班预测优化模型的主要内容,如果未能解决你的问题,请参考以下文章