内河液化天然气加注码头储罐分配与加注船调度集成优化  

Integrated planning of tank allocation and bunker ship scheduling in a liquefied natural gas bunkering terminal in the inland waterways

作　　者：刘保利 余德平 郑建风 LIU Baoli;YU Deping;ZHENG Jianfeng(Transportation Engineering College,Dalian Maritime University,Dalian 116026,China;School of Management,Huazhong University of Science and Technology,Wuhan 430074,China)

机构地区：[1]大连海事大学交通运输工程学院,大连116026 [2]华中科技大学管理学院,武汉430074

出　　处：《系统工程理论与实践》2023年第12期3668-3684,共17页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(72301051,72371130,72371046,72001194)。

摘　　要：为提高内河液化天然气加注码头运作效率,减少船舶加注等待时间,研究固定计划期内该类码头的储罐分配与加注船调度集成优化.重点考虑移动式加注与固定式加注的交互影响,兼顾储罐/加注船与受注船匹配限制、受注船可服务时间窗与最大服务时间限制、储罐库存与加注船容量限制等现实约束,以所有受注船总加权等待时间最小为目标,构建混合整数线性规划模型.基于问题的特点,重建原模型为集划分模型,并设计行列生成算法求解.通过数值实验,验证了方案及算法的有效性;特别地,行列生成算法的求解结果同下界平均相对偏差不足0.01%,且平均求解耗时少于2分钟;分析了储罐/加注船数和受注船最大服务时间的影响,可为加注码头运营提供决策支持.To improve the bunkering operation efficiency of an LNG bunkering terminal in the inland waterways,as well as reduce the waiting time for all bunker ships,this paper investigates the integrated planning of tank allocation and bunker ship scheduling in such a terminal within a fixed planning horizon.We formulate the problem as a mixed-integer linear programming model,with the objective of minimizing the total weighted waiting time of all LNG-refueled ships.Our model considers the interactions between mobile and stationary bunkering modes,along with some realistic factors such as the matching restrictions between tanks/bunker ships and LNG-refueled ships,the serviceable time windows and maximum service times of ships,the tank inventory limitations and the truck capacity constraints.Based on the characteristics of the problem,the original model is reformulated as a set partitioning model,and a row-column generation algorithm is developed to solve the reformulated model.Numerical experiments are conducted to validate the effectiveness of the proposed model and algorithm;in particular,the average relative deviation of the solution results of the row-column generation algorithm from the lower bound is less than 0.01%,and the average computational time is less than 2 minutes;the effects of the number of tanks/bunker ships and the maximum service time of the LNG-refueled ships are analyzed,which can provide decision support for operations and management at the bunkering terminal.

关 键 词：液化天然气 加注码头 储罐分配 加注船调度 行列生成算法 

分 类 号：U691.3[交通运输工程—港口、海岸及近海工程]