Liner shipping network design with emission control areas: A genetic algorithm-based approach

To curb emissions, containerized shipping lines face the traditional trade-off between cost and emissions (CO₂ and SO_x) reduction. This paper considers this element in the context of liner service design and proposes a mixed integer linear programming (MILP) model based on a multi-commodity pickup and delivery arc-flow formulation. The objective is to maximize the profit by selecting the ports to be visited, the sequence of port visit, the cargo flows between ports, as well as the number/operating speeds of vessels on each arc of the selected route. The problem also considers that Emission Control Areas (ECAs) exist in the liner network and accounts for the vessel carrying capacity. In addition to using the MILP solver of CPLEX, we develop in the paper a specific genetic algorithm (GA) based heuristic and show that it gives the possibility to reach an optimal solution when solving large size instances.     

Scheduled paratransit transport systems

In this paper we focus on ways to provide individualized services to people with mobility challenges using existing modes of public transport. We study the design of an interesting case, in which a bus operating in a public transport route may diverge from its nominal path to pick-up passengers with limited mobility and drop them off at their destination. We have modeled the design problem by a mixed integer-linear program, and we developed an exact Branch and Price approach to solve it to optimality. The proposed approach includes a labeling algorithm in which we introduced appropriate dominance rules, which do not compromise optimality. We have compared the efficiency of our approach with that of related algorithms from the literature. Furthermore, we have used the proposed approach to study key aspects of the system design problem, such as the effect of various constraints on the service level, and the tuning of the system’s parameters to address different transport environments.     

供应链订单任务分配模型及其混合遗传算法

建立了面向供应链的多产品、多订单、多时段的订单任务分配的混合整数线性规划模型.为缩短求解时间和避免大量非法染色体出现,设计了基于遗传算法和启发式规则相结合的混合遗传算法.通过算例比较了该算法和SAS软件的OR模块.结果表明,与SAS/OR模块相比,该算法最优解误差为2.3%,但计算运行时间仅为SAS/OR模块的0.89%.     

回收物流设施多层选址模型及其算法

针对废弃物品的收集、预处理和再制造问题，建立了回收物流设施选址混合整数规划（MILP）模型，用分枝定界法求解，为提高分枝定界法的收敛速度、减少运算量，先用遗传算法为MILP找到较好的可行解，以其对应的目标函数值为分枝界定法上界，从而可以在分枝界定法的搜索中剪去大量分枝．用算例证明了算法的有效性和可行性．     

基于虚拟车队的自动交叉路口车辆时序优化模型

智能网联汽车可通过彼此交互协同安全地通过交叉路口,自动交叉路口控制已成为未来发展趋势。为解决现有基于预约的自动交叉路口控制模型未全局优化车辆通过顺序及模型非线性导致求解效率低等问题,提出一种基于虚拟车队的自动交叉路口车辆时序优化模型,实现车辆通过时序的高效全局优化。首先,为构建到达安全时间间隔约束,基于车辆冲突分析计算交叉口进口道停车线到各相互作用点的距离。其次,为便于建模和求解,基于时间维度构建虚拟车队并形成车辆索引序列。然后,以交叉口车辆总延误最小为优化目标,车辆通过控制区段的最小行程时间和到达冲突区域边界的安全间隔为约束条件,构建自动交叉路口车辆通过时序非线性优化模型。在此基础上,引入0-1变量将该模型转化为混合整数线性规划模型,并基于开源求解器CBC对模型进行求解。最后,设计数值仿真试验验证模型的有效性并进行了模型的参数敏感性分析。研究结果表明:所构建模型在不同交通需求下优化效果均优于基于"先到先服务"规则的模型,车均延误和最大单车延误能够减少61.50%和39.73%;当安全间距和优化周期较大时,构建模型的延误控制效果更为显著;模型和算法为未来智能网联环境下自动交叉路口控制提供了一种可选的方法。     

基于混合整数规划的高速列车多区间节能优化研究

当今巨大的能源消耗成为高速铁路发展不可忽视的问题,节能优化显得格外重要。从列车优化操纵与区间运行时分分配两方面出发,在总运行时分约束下,探讨全程最节能的运行时分分配方案及所对应的列车操纵策略。在列车运动方程基础上,引入经停站约束,建立列车运行能耗和区间运行时分分配同步优化的非线性连续节能优化模型。通过离散化与线性变换,将节能优化模型重构为混合整数规划模型,并利用Cplex求解。以高速铁路线路为背景设计算例,验证模型的有效性。仿真结果证明在列车优化操纵策略的基础上,优化区间运行时分分配方案可以进一步降低列车总运行能耗。在不改变总运行时间的情况下,运行总能耗可降低1.03%。     

靠泊位置受限的连续泊位-岸桥同步分配模型

提出了一个连续泊位-岸桥同步分配的混合整数线性规划(MILP)模型,可以限制不同类型船舶的靠泊位置.同时,模型还考虑了岸桥干扰和靠泊位置偏离导致装卸效率损失的因素,发现可对船舶装卸时分配的岸桥数进行动态调整.利用现有文献的基准测试实例对模型进行测试,比较靠泊位置受限和不受限情况下对服务成本的影响.此外,还通过实验测试了岸桥干扰和靠泊位置偏离对总服务成本的影响,发现服务成本对这两个因素的变化非常敏感,且岸桥干扰因素的影响更大.     

石油产品管道运输定量研究

成品油管道运输计划和调度是具有多约束的复杂问题,包括炼油厂的生产进度、市场需求、相邻油品的规格等都需考虑。另外,炼油厂和分销中心现有库存量和库存能力以及管道中的输油量等也要详细分析。文中对该领域的定量研究文献进行了梳理、综述,对未来研究趋势做了分析展望,以便研究人员作进一步研究。