串行式混合类型航道船舶交通组织优化  

SHIP TRAFFIC ORGANIZATION OPTIMIZATION OF SERIAL MIXED CHANNEL

作　　者：王志强[1] 张新宇[1] 李倍莹 王婧贇 Wang Zhiqiang;Zhang Xinyu;Li Beiying;Wang Jingyun(Maritime Intelligent Transportation Research Group,Dalian Maritime University,Dalian 116026,Liaoning,China)

机构地区：[1]大连海事大学海上智能交通研究组,辽宁大连116026

出　　处：《计算机应用与软件》2023年第2期26-32,39,共8页Computer Applications and Software

基　　金：国家自然科学基金项目(51779028)。

摘　　要：随着港口航道类型逐渐向多样化、复杂化的混合类型趋势发展,港口交通问题愈发严峻。调研国内外港口混合类型航道,抽象出一种串行式简单混合类型航道作为研究对象。分析混合航道船舶交通状况,构建以单向/混合通航模式转化、混合航道异类子航道间通航模式切换、港池连接水域船舶交通冲突消解等为约束的串行式简单混合类型航道船舶交通组织优化模型。基于Spark并行计算框架,结合NSGA-II算法遗传操作天然并行性特点,提出一种Spark分布式多目标遗传算法,将全部种群分散在多节点上并行执行算法的遗传操作。实验表明,提出的算法具有较快的收敛速度和较好的稳定性,模型求解出的优化方案合理、有效。With the development of port channel type to the mixed type of diversification and complexity,port traffic problems become more and more serious.Through investigation of the mixed channels at home and abroad,we abstracted a simple serial mixed channel as the research object.Based on the analysis of the ship traffic situation in the mixed channel,we constructed ship traffic organization optimization model of a serial simple mixed channel which was constrained by single/mixed channel navigation mode transforming,different sub channels of mixed channel navigation mode switching,and ship traffic conflict resolution in the harbor connected waters,etc.Based on Spark parallel computing framework,we proposed a Spark distributed multi-objective genetic algorithm by combining with the natural parallelism of NSGA-II genetic operation.We distributed all the population on multi nodes and performed the genetic operation of the algorithm in parallel.The experiment shows that the proposed algorithm has fast convergence speed and good stability,and the optimization scheme solved by the model is reasonable and effective.

关 键 词：混合类型航道 船舶交通组织优化 分布式多目标遗传算法 Spark并行计算框架 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术] U691.32[交通运输工程—港口、海岸及近海工程]