后疫情时代基于XGBoost的铁路客运站客流量预测研究

进入"后疫情时期",铁路客流正逐步回升,但呈现较大波动,面对铁路提质增效的任务,准确预测客流量愈发重要.文章采用极端梯度提升(XGBoost,eXtreme Gradient Boosting)模型,以新冠肺炎疫情、天气和日期属性作为影响因素,选取上海站2016年1月1日—2020年7月27日客流量数据作为训练集和验证...     

基于遗传算法的天然气集输管网参数优化设计

以管道建设总费用为目标建立目标函数 ,以管道的稳态分析、各节点的流量、压力及管道压力限制等为约束条件 ,建立了天然气集输管网的参数优化设计模型。该模型属非线性离散化最优组合问题 ,采用遗传方法求解 ,并给出了计算机软件算法     

汽车灯配光镜CAD的数据结构及搜索方法

汽车灯配光镜是一种面形复杂，边界不完全规正的微透镜阵列，其数据结构及搜索方法是CAD软件实现的难点之一，本文给出一套可行的数据结构及搜索方法，经依此编制成的程序（MVHLD）对四种不同的前照灯所作的400多万条光线[1]的计算证实这套数据结构是合理的。搜集方法是可靠的。     

Equilibrium properties of taxi markets with search frictions

The equilibrium properties of an aggregate taxi market are investigated using a general bilateral searching and meeting function which characterizes the search frictions between vacant taxis and unserved customers. Three specific issues are analyzed for meeting functions that exhibit increasing, constant and decreasing returns to scale. Firstly, service quality in terms of customer wait/search time and average profit per taxi are examined jointly in relation to taxi fleet size, and a Pareto-improving win-win situation is identified, where an increase in taxi fleet size leads to improvements in both service quality and market profitability. Such a Pareto-improving situation is found to emerge if and only if the meeting functions show increasing returns to scale. Secondly, the properties of the socially optimal solution are examined. It is found that the taxi fleet size should be chosen such that the total cost of operating vacant taxis equals the total cost of customer waiting time multiplied by an asymmetric factor of the meeting function, and that taxi services should be subsidized at social optimum only when the meeting functions show increasing returns to scale. Thirdly, the Pareto-efficient services are examined for trade-offs between social welfare and profits in the light of partially conflicting objectives of the public sector and the private taxi firms using a bi-objective maximization approach. The taxi utilization rate and the customer wait/search time or service quality are proved to be constant along the Pareto frontier and equal to those at social optimum if the meeting functions show constant returns to scale. Extensions are made to the cases with increasing and decreasing returns to scale.     

基于MATLAB和TOPSIS法的齿轮齿条转向器的参数优化

通过Matlab编程,在满足理想关系式且外轮转角小于内轮转角的前提下,求出内外轮转角、主销偏移距、最小转弯半径,算出作用在方向盘上的手力、转向器的传动比、原地转向阻力、作用在转向节上的阻力。运用TOPSIS法进行评价,在主要考虑汽车转向轻便性与转向灵敏性的大条件下,同时兼顾转弯机动性的情况,作用在方向盘上的手力、转向器的传动比、最小转弯半径的权重按0.45、0.45、0.1的比例选取,最后通过具体案例确定齿轮齿条转向器角传动比、最小转弯半径、主销偏移距、外轮转角、内轮转角、转向盘最大转动圈数、作用在方向盘上的手力,为齿轮齿条转向器优化设计提供了重要的方法。     

公铁联运规模效应和枢纽拥挤效应的联动分析

公铁联运最主要的优势就是采用轴辐式网络结构实现规模经济,降低货物的运输成本,货运量越大,规模效应越明显;另一方面,过多的货物在枢纽集结会造成枢纽拥挤,增加货物在枢纽的停留时间,从而降低了公铁联运的竞争力.本文分别采用与流量相关的分段成本函数和M/M/c的排队模型来描述公铁联运的规模效应和拥挤效应,建立了非线性的整数规划模型,用于综合分析规模效应和枢纽拥挤效应对公铁联运的影响.采用混合禁忌搜索算法求解本文的模型,该算法具有较好的收敛效果.最后,分析了铁路服务固定成本、公路运输成本、货物运到期限和枢纽能力对公铁联运竞争力的影响.     

基于改进遗传算法的舾装件托盘多载具协同拣选方法

为提升舾装件托盘的拣选效率,建立拣选过程的数学模型,提出一种基于改进遗传算法（Improved Genetic Algorithm,IGA）的舾装件托盘多载具协同拣选方法。针对遗传算法（Genetic Algorithm,GA）流程与实际拣选过程的差异,改进GA的初始化过程和染色体交叉方式,并对变异过程进行更贴近实际生产的修改。针对GA难以得到全局最优解的问题,采用变邻域搜索（Variable Neighborhood Search,VNS）策略降低陷入局部最优解的可能性。采用实例计算验证该算法的有效性,可优化传统舾装件托盘拣选方法。     

给定站点条件下轨道交通线网初步辅助构建模型

为解决城市轨道交通线网规划实践中，缺乏量化为主的成网技术流程，站点数量较多情况下规划工作复杂等问题，基于给定站点，将站点间邻接关系和网络上的边是否存在线路以0-1变量表示，并作为决策变量，分别以线网总里程最小化和运营效率最大化为目标，建立由网络设计模型和线路选择模型组成的初步辅助构建模型；并在站点可达性和工程可行性的约束下，设计基于聚类的最小生成树算法和网络修正方案，求解网络设计模型，提高网络生成速度，修正不合理的网络结构；并在运营可行性约束下，设计基于禁忌算法的双目标规划求解算法，求解线路选择模型，成功将其转化为单目标问题，提高算法寻优速度。将实际运营轨网站点和客流OD作为输 入，双层模型可获得最优网络和线路方案，并与实例比较证明结果的合理性。结果表明，模型结果与已开通运营轨网近似，但线网总里程更低，运营效率更高，线网总里程降低1.7%，直达客流比增加0.7%，一次换乘客流比增加4.2%。     

A tabu search algorithm to solve the integrated planning of container on an inter-terminal network connected with a hinterland rail network

Transport demand for containers has been increasing for decades, which places pressure on road transport. As a result, rail transport is stimulated to provide better intermodal freight transport services. This paper investigates mathematical models for the planning of container movements in a port area, integrating the inter-terminal transport of containers (ITT, within the port area) with the rail freight formation and transport process (towards the hinterland). An integer linear programming model is used to formulate the container transport across operations at container terminals, the network interconnecting them, railway yards and the railway networks towards the hinterland. A tabu search algorithm is proposed to solve the problem. The practical applicability of the algorithm is tested in a realistic infrastructure case and different demand scenarios. Our results show the degree by which internal (ITT) and external (hinterland) transport processes interact, and the potential for improvement of overall operations when the integrated optimization proposed is used. Instead, if the planning of containers in the ITT system is optimized as a stand-alone problem, the railway terminals may suffer from longer delay times or additional train cancellations. When planning the transport of 4060 TEU containers within one day, the benefits of the ITT planning without considering railway operations account for 17% ITT cost reduction but 93% railway operational cost growth, while the benefits of integrating ITT and railway account for a reduction of 20% in ITT cost and 44% in railway operational costs.