港口调度算法的推荐与部署方法  

Recommendation and Deployment Methods of Port Scheduling Algorithms

作　　者：陈德显 林志诚 Chen Dexian

机构地区：[1]东南大学自动化学院,江苏南京210096

出　　处：《工业控制计算机》2025年第2期65-67,共3页Industrial Control Computer

摘　　要：为应对当前日益增长且多样化的港口设备调度需求,实现调度算法的智能选择和推荐是一种新颖的最佳解决方案之一。提出了一种基于分类回归树(CART)的调度算法的推荐方法,以及容器化的调度算法快速部署方法。首先,收集整理港口中各种多机调度需求并进行特征选取与数据预处理,然后利用决策树预剪枝技术和Sklearn库构建CART决策树。同时,利用Jenkins自动化服务器建立了容器化的调度算法的自动部署流程。实验结果表明,提出的方法在面对不同调度需求时均可准确推荐合适的算法,且算法部署快速便捷。To address the growing and diversifying demands for port equipment scheduling,intelligent selection and recommendation of scheduling algorithms emerge as one of the novel optimal solutions.This paper presents a recommendation method for scheduling algorithms based on classification and regression trees(CART)and a rapid deployment method for containerized scheduling algorithms.Initially,various multi-machine scheduling demands in ports are collected and organized for feature selection and data preprocessing.Subsequently,a CART decision tree is constructed using decision tree pre-pruning techniques and Sklearn.In parallel,an automated deployment process for containerized scheduling algorithms is established using Jenkins.Experimental results demonstrate that the proposed method can accurately recommend suitable algorithms for different scheduling demands with rapid and convenient algorithm deployment.

关 键 词：港口调度 决策树 容器化 持续集成 

分 类 号：U691.3[交通运输工程—港口、海岸及近海工程] TP181[交通运输工程—船舶与海洋工程]