Coordinated scheduling model for intermodal transit hubs based on GI/M^K/1 queuing system

Coordinated scheduling of multimode plays a pivotal role in the rapid gathering and dissipating of passengers in transport hubs. Based on the survey data, the whole-day reaching time distribution at transfer points of passengers from the dominant mode to the connecting mode was achieved. A GI/M K/1 bulk service queuing system was constituted by putting the passengers' reaching time distribution as the input and the connecting mode as the service institution. Through queuing theory, the relationship between average queuing length under steady-state and headway of the connecting mode was achieved. By putting the minimum total cost of system as optimization objective, the headway as decision variable, a coordinated scheduling model of multimode in intermodal transit hubs was established. At last, a dynamic scheduling strategy was generated to cope with the unexpected changes of the dominant mode. The instance analysis indicates that this model can significantly reduce passengers' queuing time by approximately 17% with no apparently increase in departure frequency, which provides a useful solution for the coordinated scheduling of different transport modes in hubs.

Deep Reinforcement Learning Solves Job-shop Scheduling Problems

To solve the sparse reward problem of job-shop scheduling by deep reinforcement learning,a deep reinforcement learning framework considering sparse reward problem is proposed.The job shop scheduling problem is transformed into Markov decision process,and six state features are designed to improve the state feature representation by using two-way scheduling method,including four state features that distinguish the optimal action and two state features that are related to the learning goal.An extended variant of graph isomorphic network GIN++is used to encode disjunction graphs to improve the performance and generalization ability of the model.Through iterative greedy algorithm,random strategy is generated as the initial strategy,and the action with the maximum information gain is selected to expand it to optimize the exploration ability of Actor-Critic algorithm.Through validation of the trained policy model on multiple public test data sets and comparison with other advanced DRL methods and scheduling rules,the proposed method reduces the minimum average gap by 3.49%,5.31%and 4.16%,respectively,compared with the priority rule-based method,and 5.34%compared with the learning-based method.11.97%and 5.02%,effectively improving the accuracy of DRL to solve the approximate solution of JSSP minimum completion time.

Implementation and Evaluation of Dynamically Weighted Low Complexity Fair Queuing(DWLC-FQ) Algorithm for Packet Scheduling in WiMAX Networks

Services provided by internet need guaranteed network performance. Efficient packet queuing and scheduling schemes play key role in achieving this. Internet engineering task force(IETF) has proposed Differentiated Services(Diff Serv) architecture for IP network which is based on classifying packets in to different service classes and scheduling them. Scheduling schemes of today's wireless broadband networks work on service differentiation. In this paper, we present a novel packet queue scheduling algorithm called dynamically weighted low complexity fair queuing(DWLC-FQ) which is an improvement over weighted fair queuing(WFQ) and worstcase fair weighted fair queuing+(WF2Q+). The proposed algorithm incorporates dynamic weight adjustment mechanism to cope with dynamics of data traffic such as burst and overload. It also reduces complexity associated with virtual time update and hence makes it suitable for high speed networks. Simulation results of proposed packet scheduling scheme demonstrate improvement in delay and drop rate performance for constant bit rate and video applications with very little or negligible impact on fairness.

Neural Network Based Scheduling for Variable-Length Packets in Gigabit Router with Crossbar Switch Fabric and Input Queuing

A high-speed and effective packet scheduling method is crucial to the performance of Gigabit routers. The paper studies the variable-length packet scheduling problem in Gigabit router with crossbar switch fabric and input queuing, and a scheduling method based on neural network is proposed. For the proposed method, a scheduling system structure fit for the variable-length packet case is presented first, then some rules for scheduling are given. At last, an optimal scheduling method using Hopfield neural network is proposed based on the rules. Furthermore, the paper discusses that the proposed method can be realized by hardware circuit. The simulation result shows the effectiveness of the proposed method.

Distributed fair queuing algorithm based on compensation coordination scheduling in WMN

This paper proposes a distributed fair queuing algorithm which is based on compensation coordi- nation scheduling in wireless mesh networks, considering such problems as the location-dependent competition and unfair channel bandwidth allocation among nodes. The data communication process requiring the establishment of compensation coordination scheduling model is divided into three periods： the sending period, the compensation period and the dormancy period. According to model parameters, time constraint functions are designed to limit the execution length of each period. The algorithms guarantee that the nodes complete fair transmission of network packets together in accordance with the fixed coordination scheduling rule of the model. Simulations and analysis demonstrate the effectiveness of the proposed algorithm in network throughput and fairness.

Wind Turbine Optimal Preventive Maintenance Scheduling Using Fibonacci Search and Genetic Algorithm

Maintenance scheduling is essential and crucial for wind turbines (WTs) to avoid breakdowns andreduce maintenance costs. Many maintenance models have been developed for WTs’ maintenance planning, suchas corrective, preventive, and predictive maintenance. Due to communities’ dependence on WTs for electricityneeds, preventive maintenance is the most widely used method for maintenance scheduling. The downside tousing this approach is that preventive maintenance (PM) is often done in fixed intervals, which is inefficient. In thispaper, a more detailed maintenance plan for a 2 MW WT has been developed. The paper’s focus is to minimize aWT’s maintenance cost based on a WT’s reliability model. This study uses a two-layer optimization framework:Fibonacci and genetic algorithm. The first layer in the optimization method (Fibonacci) finds the optimal numberof PM required for the system. In the second layer, the optimal times for preventative maintenance and optimalcomponents to maintain have been determined to minimize maintenance costs. The Monte Carlo simulationestimates WT component failure times using their lifetime distributions from the reliability model. The estimatedfailure times are then used to determine the overall corrective and PM costs during the system’s lifetime. Finally,an optimal PM schedule is proposed for a 2 MW WT using the presented method. The method used in this papercan be expanded to a wind farm or similar engineering systems.

考虑随机返工与紧急插单的可抢占设计项目群调度

鉴于多项目并行以及因随机返工和紧急插单而导致的任务抢占特征,使设计项目群调度问题难以求解,而传统随机动态规划方法因维数灾难只能获得小规模问题的最优策略,为构造面向大规模问题的高效近似方法,提出将原随机调度问题近似分解为各个状态下的确定性调度子问题,从而获得原问题的次优策略。建立了确定性调度子问题的混合整数规划模型,并提出基于元启发式算法和优先规则的求解方法。基于标准测试集PSPLIB设计了多组计算实验,在不同调度环境下验证了所构建模型和算法的有效性。实验结果表明,元启发式算法不但在单项目平均拖期比目标下较最优规则提升了12%以上,而且具有较高的计算效率,能够满足实际决策需求。

基于排队论的一种高吞吐量信道分配协议设计

针对复杂多变的认知无线电系统中难以为次用户高效分配信道的问题,提出了一种高吞吐量信道分配协议——TKMA协议。该协议根据主用户(PU)活动、次用户(SU)实时业务需求、信道条件等信息构建用户信道的效用矩阵,在保障PU通信质量的前提下以SU系统总效用值最大化为目标进行信道分配,并利用改进Kuhn-Munkras算法结合轮询调度进行求解。为了评估该协议性能,建立了通用的多用户多信道认知无线电系统模型,利用排队理论描述数据包传输过程,并通过马尔可夫稳态求解推导出SU的性能指标。实验结果表明,与以往提出的化简方法和传统的公平随机分配协议相比,使用TKMA协议在SU系统总的吞吐量、平均时延、平均队长、拒绝率等指标上都取得了更优的结果,证明了所提协议和系统模型的有效性。

混合数据的多集群系统中数据价值与信息年龄的联合优化

信息年龄(AoI)是一种业界新兴的时间相关指标,其经常用于评估接收数据的新鲜度。该文考虑了一个视频数据与环境数据混合的多集群视频直播系统,并制定调度策略以联合优化系统数据价值与信息年龄。为克服优化问题中动作空间过大导致难以实现有效求解的问题,该文将优化问题的调度策略分解为相互关联的内外两层策略,外层策略利用深度强化学习实现集群间的信道分配,内层策略则基于构造的虚拟队列实现集群内的链路选择。双层调度策略将每个集群的内层策略嵌入到外层策略中进行训练,仿真结果显示,与现有调度策略相比,该文所提的调度策略可以提高时间平均的接收数据价值并降低时间平均的信息年龄。

计及灵活性资源的综合能源系统源荷协调优化调度

可再生能源和负荷的波动性、不确定性等给综合能源系统(integrated energy system,IES)的安全灵活运行带来了极大挑战。为提高IES灵活调节能力与可再生能源消纳水平,提出一种计及灵活性资源的IES源荷协调优化调度方法。针对系统内运行灵活性需求,精细刻画各类资源灵活性能力,源侧根据电氢耦合单元运行特性构建热电联产机组(combined heating and power,CHP)和氢燃料电池(hydrogen fuel cell,HFC)联合运行模型,荷侧考虑综合需求响应的灵活性供给能力,建立系统综合灵活性供给模型。根据不同时刻运行灵活性不足问题分成2种调度模式,构建基于IES灵活性约束的优化调度模型,并进行仿真分析。仿真结果表明,所提出的优化调度方法能够有效提高IES灵活调节能力和可再生能源消纳水平。

带有N-策略和负顾客的可修重试排队系统均衡策略研究

本文研究带有N-策略和负顾客的M/M/1可修常数重试排队系统中正顾客的策略行为.当系统中的正顾客数达到规定阈值N时,处于空闲状态的服务台才会重新启动提供服务.系统中没有等待空间,若正顾客在服务台处于空闲状态时抵达,则他会立即接受服务;若正顾客在服务台处于忙期时抵达,则他会进入重试轨道,或者选择离开系统.当服务台再次为空时,它会遵循先到先服务的规则从重试轨道中选择正顾客进行服务.负顾客在到达系统时会抵消正在被服务的正顾客,并且导致服务台发生故障.若服务台发生故障,它会立刻被送去修理,修理时间服从指数分布.我们首先推导出系统的稳态概率和正顾客的平均逗留时间,求得了不同状态下正顾客的均衡到达率和单位时间社会收益.最后,对正顾客的社会最优到达率和最优社会收益进行了数值分析.

考虑优先级的智能电网业务调度与资源分配方案

随着智能电网和5G通信技术的融合发展,越来越多的智能终端应用到智能电网系统。针对海量电力业务的分流处理问题,提出了一种考虑电力业务优先级的业务调度和资源分配方案。首先介绍了面向智能电网的基于软件定义网络的边缘计算处理架构,建立了业务处理模型。其次阐述了基于强占型优先级排队的业务调度机制,建立了业务卸载收益和卸载开销的数学模型,该模型以系统整体收益最大化为目标函数,并基于电力业务的卸载有效性得到每一种优先级业务在边缘服务器的资源分配阈值,并以资源分配阈值为约束条件。再次,利用改进的遗传算法(improved genetic algorithm,IGA)求解最优的卸载和资源分配决策。最后通过实验仿真验证了IGA在收敛速度和个体选择方面均优于其他对比算法,对比其他方法所提方案在业务平均处理时间、功耗、高优先级业务平均处理时间等方面分别降低了69.2%、67.7%、73%,在系统收益方面提升了119%。

基于代理双向竞拍的弹性云资源调度方案

针对跨云平台的资源调度问题,提出一种基于代理双向竞拍的弹性资源调度方案。该方案在拍卖过程中引入交易延时排队影响因子,排队轮数越多,排队影响因子值越大,竞价提交优先级越高,同时根据服务等级协议判定不合规交易的惩罚金数额,以惩罚金制度防范恶意竞拍交易的出现。每轮竞拍允许并发实现多项交易。仿真结果表明,该方案保证了竞拍成功率,并且与传统一次拍卖成交方案相比,其竞拍时间减少了0.32 s,同时能提高16%的用户收益与9%的资源提供商利益。

面向多目标状态感知的自适应云边协同调度研究

随着智能城市和工业智能制造的蓬勃发展,从监控摄像头获取详尽信息以进行多目标视觉分析的需求日益突出。现有研究主要关注在服务器上进行资源调度以及改良视觉模型,往往不能很好地应对设备资源状态和任务状态的动态变化。随着边缘端硬件资源的升级和任务处理模型的改进,设计一个自适应的云边协同调度模型来满足任务的实时用户需求成为优化多目标状态感知任务的重要方式。因此,在深入分析云边场景下多目标状态感知任务特性的基础上,提出了一种基于深度强化学习的自适应云边协同调度模型ATS-SAC。ATS-SAC通过实时解析多目标状态感知任务的运行时状态,动态给出任务执行的视频流配置、模型部署配置等调度决策,从而显著优化环境不稳定的云边场景下多目标状态感知任务的精度、时延的综合性能质量。同时,还引入了一种基于用户体验极限阈值的动作筛选方法,有助于去除冗余的决策动作,进一步优化模型的决策空间。针对用户对多目标状态感知任务性能结果的不同需求,ATS-SAC模型能提供包括极速模式、均衡模式和精度模式在内的多种灵活的调度策略。实验结果表明,相比其他的任务执行方式,在ATS-SAC模型的调度策略下,多目标状态感知任务在精度质量和处理时延上更能满足用户的体验需求。同时,当实时运行状态变化时,ATS-SAC模型能够动态调整其调度策略,以保持稳定的任务处理效果。

到达时间服从泊松分布的平行机队列的最优随机排序问题

论文考虑多个分布下,根据每类加工时间函数最小化目标函数的不同类别的随机排序问题。这个问题常出现在分布式系统、网络和应用程序方面。模型中,最优排序策略在每台机器上是一个简单的静态优先策略。在这种排序策略下,排序问题可以寻找到最佳路径矩阵。考虑一个非线性规划问题,证明了任何局部最优即为全局最优,大大简化了,优化问题的解决方案。在到达时间为泊松分布的情形下,论文提供了一个最佳的排序策略,能够最小化每类时间函数。对一般各种静态实例应用此方法,可得到简单的近似算法。

基于CNN的深度强化学习算法求解柔性作业车间调度问题

在使用深度强化学习(DRL)算法解决柔性作业车间调度(FJSP)问题时,状态和动作的表示具有复杂多变的特性,这导致算法的求解质量不高,为了得到更优解,对状态和动作的表示进一步研究,以最大完工时间最短为优化目标,采用卷积神经网络(CNN)和近端策略优化(PPO)方法设计了一种深度强化学习算法。针对柔性作业车间环境的复杂性,特别设计了双通道状态表示方法,第一通道表示每道工序选择的机器,第二通道表示每道工序在所选机器上的加工次序。在动作设置中设计了一种机器选择算法,能够根据当前状态选取最佳的机器,搭配深度强化学习算法共同组成动作的选择。通过Brandimarte算例验证表明,该算法具有可行性,比常用深度强化学习算法求解质量更优,在不同规模算例的性能表现更好。

基于策略梯度的WRSN公平能量补充方案

无线可充电传感网络(WRSN)中节点能耗是动态变化的,在调度移动充电器(MC)为传感器充电时,估算节点的实时能耗来计算节点的剩余生存时间十分重要。针对WRSN的问题,提出一种基于策略梯度的公平能量补充策略(PFSS),利用过往的节点能量记录对节点实时能耗进行估算,采用seq2seq结合Attention的网络结构求解待充电序列,为节点充电获得一个充电权重来评估该次充电的质量,其由MC到节点间的距离、节点等待时间和节点剩余生存时间组成,利用策略梯度对网络进行训练,从而求得具有最优充电质量的充电序列近似解,仿真结果表明,与现有的算法相比,PFSS能够有效降低网络中传感器节点失效数量、MC的移动成本和充电延迟。

基于DDPG算法的离网型可再生能源大规模制氢系统优化调度

为提高可再生能源消纳能力,减少整流和并网等设备的投资成本,降低电解水制氢系统成本,实现可再生能源大规模制氢,构建了一个孤岛型可再生能源大规模制氢系统。该系统通过智慧能量管理,实现了提高系统经济性与安全性的目标。首先建立可再生能源大规模制氢系统的仿真模型,制定控制策略;其次,提出一种基于深度确定性策略梯度(DDPG)的能量优化调度策略。通过大量长期的训练,使用DDPG算法得到的智能体能够实现智能化的动态能量优化调度。将该策略与深度Q网络、粒子群优化和传统控制方法在经济性和安全性方面进行比较,结果表明DDPG算法在能量优化管理中可实现更高的经济收益,更好地利用可再生资源,并确保系统的安全运行。

立体车库出入库路径规划优化算法

目前,伴随着中国的发展,人们的生活质量也有了显著的提升,我国的私人汽车也越来越多。随着我国城镇私人汽车拥有量的不断增加,我国城镇居民的“停车难”问题已经成为当前急需解决的一大难题,也是当前我国城镇居民面临的迫切需求。当前,国内的立体停车设施普遍面临着停车周期长,运营效率低,维护成本高等问题,使其很难在中小城市得到普及。造成这一问题的主要原因是,立体停车场的自动测量和控制是一项集成了机电、光学和计算等多个领域的交叉技术,它的研究和实施都有很大的困难。在这些问题中,最重要的是对立体停车场进行有效的调控和优化。另外,由于我国中小城镇居民对立体停车设施的认可度较差,使得整个市场出现了“有价无市”的局面。为了改善和解决在现阶段,在三维车库的测控管理中存在的上述问题,本文就是要对三维车库的测控方法和优化策略进行讨论和分析,重点是对排队模型的构建和路线优化算法的实现进行研究。

Study on scheduling algorithm for multiple handling requests of single automated guided vehicles

To solve the problem of small amount of machining centers in small and medium flexible manufacture systems(FMS), a scheduling mode of single automated guided vehicle(AGV) is adopted to deal with multiple transport requests in this paper. Firstly, a workshop scheduling mechanism of AGV is analyzed and a mathematical model is established using Genetic Algorithm. According to several sets of transport priority of AGV, processes of FMS are encoded, and fitness function, selection, crossover, and variation methods are designed. The transport priority which has the least impact on scheduling results is determined based on the simulation analysis of Genetic Algorithm, and the makespan, the longest waiting time, and optimal route of the car are calculated. According to the actual processing situation of the workshop, feasibility of this method is verified successfully to provide an effective solution to the scheduling problem of single AGV.