DIAGNOSTIC CHECKING FOR TIME SERIES MODELS USING NONPARAMETRIC APPROACH

提出一种时间序列模型残差诊断捡验的非参数方法．基于残差与时间ｔ的最大相关系数ρ２＊，在零假设Ｈ０∶ρ２＊＝０下，用Ｂｏｏｔｓｔｒａｐ方法建立统计量ρ２＊的零分布。为计算ρ２＊，提出了一种与ＡＣＥ算法相似的ρ算法。有效性分析表明该方法比Ｌｊｕｎｇ－Ｂｏｘ检验更有效。同时还列出一计算机模拟结果和两个实例．

Timed-Automata Based Model-Checking of a Multi-Agent System: A Case Study

A multi-agent based transport system is modeled by timed automata model extended with clock variables. The correctness properties of safety and liveness of this model are verified by timed automata based UPPAAL. Agents have a degree of control on their own actions, have their own threads of control, and under some circumstances they are also able to take decisions. Therefore they are autonomous. The multi-agent system is modeled as a network of timed automata based agents supported by clock variables. The representation of agent requirements based on mathematics is helpful in precise and unambiguous specifications, thereby ensuring correctness. This formal representation of requirements provides a way for logical reasoning about the artifacts produced. We can be systematic and precise in assessing correctness by rigorously specifying the functional requirements.

基于Time Petri Nets的实时系统资源冲突检测

Time Petri Nets在实时系统的建模和性能分析中得到广泛应用,而冲突是Petri网及其扩展模型的重要行为,解决冲突是正确分析模型动态行为的关键.目前随机Petri网、混合Petri网和区间速率连续Petri网的冲突检测方法由于没有考虑到时间约束因此无法在TPN网中使用.时间约束的引入使得Time Petri Nets模型的使能和触发语义比Petri网模型的语义复杂,冲突检测变得更加困难.为了计算冲突发生的时间和概率,首先根据时间约束,给出了变迁持续使能时延迟区间的计算方法,并证明了该方法的合理性和完备性;然后在此基础上定义并证明了Time Petri Nets模型中不冲突的检测方法;并提出了Time Petri Nets模型的冲突检测方法,给出了冲突时间区间和变迁实施概率的计算方法;最后通过实例验证说明了该方法的正确性和有效性.

“一站式”检查智能预约平台的设计与实现

为解决门诊病人检查、预约难的问题,改善患者就医体验。该文借鉴流程优化、业务重组及分时段预约的理念和方法,提出“一站式”检查智能预约平台的构建。该平台的应用,系统能自动完成预约60%的门诊检查项目,患者减少了2次排队,检查等候时间平均缩短50 min。“一站式”检查智能预约平台的搭建,真正实现让患者“看病少排队”、“检查少跑腿”,在创造良好的就医环境、提高医院服务品质等方面具有重要意义。

基于导航卫星的高精度授时技术

基于导航卫星授时长期稳定度高的特点,结合铷原子钟优良的短期稳定度和低漂移率特性,提出了一种高精度授时技术。以导航接收机1 PPS为基准对铷原子钟进行驯服,通过高精度时差测量、卡尔曼滤波、增量式PID算法3个关键技术对外提供高精度授时服务。验证结果表明,输出1 PPS相位稳定性达到±0.2 ns以内,驯服期间1 s稳定度达到3.76E-12。

基于Time Petri Net的实时系统冲撞检测与消解

time Petri net(TPN)在实时控制系统的建模中得到广泛应用,而冲撞是Petri网及其扩展模型的重要行为,解决冲撞是正确分析模型动态行为的关键.由于引入时间约束,使得TPN模型的使能和触发语义比Petri网模型的语义复杂,冲撞的检测及消解变得更加困难.首先根据时间约束,给出了变迁持续使能时延迟区间的计算方法,并证明了该方法的正确性;然后在此基础上定义并证明了TPN模型中冲撞的检测方法;给出了冲撞时间区间及修改时间约束的冲撞消解方法;最后通过实例验证说明了该方法的有效性和正确性.

电网稳定控制策略在线校核系统设计与应用

随着电力市场方式多变,新能源占比逐步提高,原有离线策略能否满足电网实时运行需要进一步验证。首先研究稳控策略在线校核系统软硬件架构,采用可伸缩灵活分布式架构,方便模型和计算数据共享同步;研究稳控策略在线校核系统对外接口,从安稳主站获取安控站当值策略信息,从调度主站获取电网实时运行方式信息,从离线分析系统获取电网元件参数信息;研究策略在线校核总体计算流程,系统解析当值策略文件后,调用电气拓扑服务程序和名称匹配程序完成电网参数提取和实时计算拓扑生成,计算服务程序再根据当值策略中预想故障分析在实时方式中策略动作后的电网稳定性,将不稳定结果通过智能告警接口推送至调度主站,为调度人员提供辅助决策。

基于InSAR的北京平原区地面沉降与地下水位演化关联特征迭代奇异谱分析

北京平原区是地面沉降发展严重的区域之一,分析其地面沉降的时空演化规律可为城市地面沉降有效防控提供数据支撑。该文利用PS-InSAR方法获取地面沉降信息,并用水准监测数据验证其精度,然后基于奇异谱分析和频谱检验提出迭代奇异谱分析(Iterative Singular Spectral Analysis,ISSA)方法,并基于该方法将地面沉降的长时序数据分解为主趋势、周期特征,得到研究区地面沉降和地下水位的时序演化特征:①研究区地面沉降的主趋势特征表现为2011—2016年沉降持续、沉降减缓和再度发展3个阶段,2017—2020年沉降持续发展,但总体趋势较稳定。②研究区地面沉降季节性差异明显,且2011—2020年夏季地面沉降比冬季严重。③地下水超采是北京平原区地面沉降的主因,在严重沉降区和较严重沉降区,第二、三承压含水层的水位变化与地面沉降呈强正相关性。

一种平面密封单向阀低温环境下开启性能优化

某平面密封单向阀在常温环境下的开启压力不大于0.25 MPa时,密封性和流量均满足要求,但在低温-55℃,保温2 h后,在低温状态下通入0.5 MPa压力的氮气,活门不能瞬间打开,延迟1~2 min才缓慢开启。低温环境下单向阀的开启性能不稳定,开启时间延迟等问题亟须解决,若开启时间较长,会影响整个系统的各项性能,造成严重的质量事故。因此,必须从产品的结构设计、工作原理和环境条件等方面进行分析,找到低温环境下引起产品开启时间延迟的主要原因,并采取有效措施,提高单向阀低温环境下的开启性能。

融合用户活跃度的上下文感知兴趣点推荐算法

为了提高兴趣点(point of interest,POI)推荐的准确性和个性化,提升用户对推荐结果的满意度,针对不同活跃度用户的特点,提出一种融合用户活跃度的上下文感知兴趣点推荐算法(A POI recommendation algorithm that integrates geographical,categorical,and temporal factors,while simultaneously considering user activity),简称AU-GCTRS。首先,为缓解数据稀疏性和冷启动问题,引入多维上下文信息;其次,通过挖掘用户签到频率、签到兴趣点数量和签到时间,将用户划分为不同活跃度的群体;最后,综合用户活跃度与上下文分数,将得分高的前K个兴趣点推荐给用户。在真实数据集上进行实验表明,AU-GCTRS算法比其他流行算法更有效地缓解了数据稀疏性和冷启动问题,提高了推荐准确率和召回率。

悬索桥主缆时变可靠度研究

悬索桥服役期内由于长期暴露在潮湿的大气环境中,极易受到大气、风雨的侵蚀,导致主缆承载力发生退化,会严重影响桥梁运营安全,需要评估腐蚀后主缆的可靠性。文章以某悬索桥为工程实例,建立了Midas Civil有限元计算模型,得到了主缆控制截面恒载和车辆荷载的荷载效应。根据主缆失效模式建立了极限状态方程,计算了主缆抗力年平均损失率为0.2%~0.4%时主缆控制截面的时变可靠度指标。研究表明,当主缆抗力退化较快时,运营80 a主缆的可靠度指标低于规范值,在设计时应采取防护措施降低运营期间主缆的腐蚀程度,以延长主缆的使用寿命。

品管圈活动在提高手术“Time-Out”执行率中的应用

目的探讨品管圈活动对手术"Time-Out"执行率的影响,以预防错误手术的发生,提高手术室管理的工作质量。方法我院2014年开始在手术室应用品质管理活动圈(品管圈)来提高手术"Time-Out"执行率。选择2013年在我院手术室行剖宫产手术的3798例为对照组,2014年品管圈活动实施后的4081例为研究组,比较两组手术"Time-Out"执行率情况、手术不良事件的发生率、患者的满意度、麻醉医师的满意度、手术医生的满意度、手术护士的满意度等。结果 (1)研究组中"Time-Out"执行率、患者的满意度、麻醉医师的满意度、手术医师的满意度及手术护士的满意度均明显高于对照组,差异有统计学意义(P<0.05);(2)研究组中手术不良事件的发生率明显低于对照组,差异有统计学意义(P<0.05)。结论手术室中开展品管圈活动不仅能提高手术"Time-Out"执行率,降低手术不良事件的发生率,提高患者及手术相关医务人员的满意度及提高手术室管理的工作质量,在临床应用中值得推广。

脊柱手术患者Time-out前期病房安全核查的实施

目的有效确认脊柱手术患者术前相关信息,保证手术如期进行。方法将脊柱外科择期手术患者180例,根据入院先后顺序分为对照组(82例)和干预组(98例),对照组按常规完成术前准备,术前1d对手术患者术前相关信息进行核查,干预组在对照组基础上按自制的Time-out前期病房安全核查要素对手术所需准备项目进行核查,比较两组手术如期进行率及手术安排满意度。结果干预组手术如期进行率、手术安排满意度显著高于对照组(P<0.05,P<0.01)。结论对脊柱手术患者实施Time-out前期病房安全核查,可有效确认患者术前相关核查的要素,充分发挥手术安全监测、预警作用,保证术前准备到位,利于提高科室整体医疗质量。

White Noise Analysis: A Measure of Time Series Model Adequacy

The purpose of this study is to apply white noise process in measuring model adequacy targeted at confirming the assumption of independence. This ensures that no autocorrelation exists in any time series under consideration, and that the autoregressive integrated moving average (ARIMA) model entertained is able to capture the linear structure in such series. The study explored the share price series of Union bank of Nigeria, Unity bank, and Wema bank obtained from Nigerian Stock Exchange from January 3, 2006 to November 24, 2016 comprising 2690 observations. ARIMA models were used to model the linear dependence in the data while autocorrelation function (ACF), partial autocorrelation function (PACF), and Ljung-Box test were applied in checking the adequacy of the selected models. The findings revealed that ARIMA(1,1,0) model adequately captured the linear dependence in the return series of both Union and Unity banks while ARIMA(2,1,0) model was sufficient for that of Wema bank. Also, evidence from ACF, PACF and Ljung-Box test revealed that the residual series of the fitted models were white noise, thus satisfying the conditions for stationarity.

基于大流量止回阀实验系统的快速预测模型

针对传感器在数据获取中的局限性和无法用于对实验系统进行全面的数据收集问题,对大流量止回阀实验系统的快速预测模型技术进行了研究,建立了实验系统的快速预测模型,进行了快速预测模型的结果分析。首先,搭建了实体模型,根据大流量止回阀实验系统的结构,结合实验系统的工作原理对其进行了简化,并进行了有限元分析;然后,利用快速预测模型的关键技术构建了实验系统数据库,进行了实验系统的样本采集;通过比对不同机器学习算法的预测准确率,选择了随机森林算法,并建立了实验系统内压与应力应变的关系;最后,进行了快速预测模型的结果分析,并完成了实验系统整体预测实验和实验系统部件单独预测实验。研究结果表明:采用随机森林(RF)算法建立的快速预测模型,拟合优度(R 2)达到了0.99,相较于深度神经网络(DNN)算法和梯度提升树(GBDT)算法,拟合优度(R 2)提高了68.97%和51.47%。实验系统整体预测与实验系统部件单独预测的对比试验结果表明:整体预测模型的预测速度更快,且精度可以达到97.43%。

Development of a Model for International Traveler’s Check-In Process Using Arena Software Tool

The increasing demand for airline services amidst limited resources results in passenger dissatisfaction and dwindling revenue for airports and airlines. The dynamics of service demand and resource supply results in check-in issues for the stakeholders in the commercial aviation industry. This has the effect of impacting negatively on service performance, cost of operations, customer satisfaction, and overall throughput. Hence, this paper modelled the travelers’ check-in process at the “D Wing” of the Departure Section of Murtala Muhammed International Airport (MMIA), Ikeja, using Arena Software Tool. The work was carried out by determining the parameters of the queues at designated service points in the check-in process. The primary data required to develop the model were acquired by direct observation of passenger flow and oral interview. Thus, the average check-in time was determined. Thereafter, a model of the international check-in system of the MMIA was developed using Arena software in combination with Microsoft Office tools. The data collected were therefore inputted into the model and simulated;the real result was compared with the simulation result of 133 completions and there was no significant difference. The result showed that the model is a representation of the real system under study. Further work will be tailored towards simulation (i.e. the model will be subjected to experimentation in order to have different scenario).

Simulation of a Model for International Traveler’s Check-In Process Using Arena Software Tool

With the current situation of insecurity in Nigeria and the worsening condition of Nigerian roads, there is increasing demand for air travel. This increasing demand for airline services amidst limited resources results in passenger dissatisfaction and reduction of revenue for airports and airlines. The dynamics of service demand and resource supply result in check-in issues for all the stakeholders in the commercial aviation industry. Hence, this research simulated the developed model for travelers’ check-in process at the “D Wing” of the Departure Section of Murtala Muhammed International Airport (MMIA), Ikeja, using Arena Software Tool. The simulation of the developed model was carried out by varying five different configurations of the servers with respect to the baggage weighing machines and passenger profiling devices to obtain the minimum average check-in time (ACT) for the process, with each configuration undergoing 133 completions of simulation runs. The result of the experimentation revealed that the 3 × 3 × 4 configuration of servers produced the smallest ACT of 18.25 minutes. This translates to a difference of about 58 minutes from the 76.16 minutes ACT of the real system;representing about 76% improvement in the check-in time of passengers. This study contributed to knowledge by revealing that the main choke points in the MMIA check-in system occur at the baggage weighing and passenger profiling areas as against the check-in counter sub-section as advanced by previous works. Furthermore, the research added value to knowledge by creating a balance between customer satisfaction and cost of operations thereby accommodating the interests of the passenger and the operator who are the two main stakeholders in the commercial aviation industry.

UML statechart based rigorous modeling of real-time system

Rigorous modeling could ensure correctness and could verify a reduced cost in embedded real-time system development for models. Software methods are needed for rigorous modeling of embedded real-time systems. PVS is a formal method with precise syntax and semantics defined. System modeled by PVS specification could be verified by tools. Combining the widely used UML with PVS, this paper provides a novel modeling and verification approach for embedded real-time systems. In this approach, we provide 1) a time-extended UML statechart for modeling dynamic behavior of an embedded real-time system; 2) an approach to capture timed automata based semantics from a timed statechart; and 3) an algorithm to generate a finite state model expressed in PVS specification for model checking. The benefits of our approach include flexibility and user friendliness in modeling, extendability in formalization and verification content, and better performance. Time constraints are modeled and verified and is a highlight of this paper.

Rigorous Modeling of Real-time System Based on UML and PVS

Rigorous modeling could improve the correctness and reduce cost in embedded real-time system development for models could be verified. Tools are needed for rigorous modeling of embedded real-time system. UML is an industrial standard modeling language which provides a powerful expressi-veness, intuitive and easy to use interface to model. UML is widely accepted by software developer. However, for lack of precisely defined semantics, especially on the dynamic diagrams, UML model is hard to be verified. PVS is a general formal method which provides a high-order logic specification language and integrated with model checking and theorem proving tools. Combining the widely used UML with PVS, this paper provides a novel modeling and verification approach for embedded real-time system. In this approach, we provide 1) a timed extended UML statechart for modeling dynamic behavior of an embedded real-time system; 2) an approach to capture timed automata based semantics from timed statechart; and 3) an algorithm to generate a finite state model expressed in PVS specification for model checking. The benefits of our approach include flexible and friendly in modeling, extendable in forma-lization and verification content, and better performance. Time constraints are modeled and verified and it’s a highlight of this paper.