主机选型的多目标规划方法研究及计算机实现

本文提出了船舶柴油机选型的非线性多目标规划数学模型和介绍了基于计算机求解该数学问题的方法。为了帮助设计师选出多目标意义下的最佳推进主机,还编制了相应的计算机程序。这项研究使设计师能够借助计算机考虑众多的技术经济因素,并且把主机选型决策建立在多指标的定量分析基础上。     

机桨设计工况的最优选择

船舶航行条件经常变化,使主机和螺旋桨偏离设计的匹配工作点,从而导致机桨技术经济性能下降,在某些情况下,所谓的设计工况徒有虚名。本文分析了上述情况产生的原因,提出了解决这些问题的方法,并试图用最优化理论处理这些问题,从而将主机与螺旋桨匹配点的选择建立在最优化计算的基础上。     

Opportunistic preventive maintenance optimization for multi-unit series systems with combing multi-preventive maintenance techniques

This paper proposes a dynamic opportunistic preventive maintenance (PM) optimization policy for multi-unit series systems by integrating multi PM techniques. Two PM techniques, periodic PM and sequential PM, are considered. Whenever one of the units reaches its reliability threshold, a PM action has to be performed on that unit. At that time the whole system has to be stopped and PM opportunities arise for the other units of the system. An optimal PM practice is determined by maximizing the short-term cumulative opportunistic maintenance (OM) cost savings for the whole system. Numerical examples are given to show how this approach works. Finally, a comparison between the proposed PM policy and the other policies is given.     

带软时间窗的集货与送货多车辆路径问题节约算法

研究了物流配送中多车运输的集货与送货车辆路径规划问题,以增加时间惩罚费用的方式插入软时间窗约束,将租车费用、货车运输费用和时间惩罚费用三者之和最小作为优化目标,建立数学模型。采用启发式节约算法求解该模型,考虑时间惩罚费用和运输费用,比较每一配送节点上直接送货和间接送货的节约费用关系,求出最优配送路径。试验结果表明:当配送次数达到50次时,货车平均装载率仍能达到80%以上,该节约算法能减少货车空程行驶和租车次数,优化了全局费用。     

集装箱码头泊位-堆场-闸口的周期协同分配

为提高集装箱进出口码头在周期性环境下的作业效率,对集装箱码头泊位-堆场-闸口的周期协同分配问题进行了研究. 首先考虑泊位、堆场、闸口3类资源对船分配过程中的可用量约束、相关性约束和周期性约束,以最小化船舶总在港时间为目标函数,建立集成调度的混合整数规划模型;在此基础上设计自适应遗传算法进行求解,其上层对船舶优先级构成的编码空间展开进化搜索,下层利用启发式将优先级解码为多资源协同分配计划,并将其评价值返回至上层迭代. 数值实验显示,协同考虑泊位、堆场和闸口3类资源的集成调度,相较于传统的两阶段调度,周期计划下的船舶总在港时间缩短约20%.      

Heuristics for the identical machine scheduling problem with preventive maintenances

In this paper, two mixed integer programming models integrating production scheduling and preventive maintenances are proposed to derive the optimal solutions for the identical machine scheduling problem with unavailability constraints. In the first model, the maintenance activities are performed periodically and the objective is to minimize the makespan. In the second model, the maintenance activities are flexible and the machines’ continuous working time cannot exceed a maximum allowed time T; the objective is to minimize the total completion time of jobs. For the first problem, we propose a heuristic longest batch time (LBT) and prove that the worst case error bound of LBT is 2. For the second problem, we develop a heuristic modified smallest processing time (MSPT) based on some properties of the optimal solutions. Computational experiments show that both of the heuristics are effective and efficient compared with the results obtained by CPLEX and the other algorithms.