A kind of crowded bottleneck identification method of rail network based on seepage theoryTechnical field
The present invention relates to track operation management and big data analysis field, more particularly to a kind of rail based on seepage theoryRoad network congestion bottleneck identification method, available for support track operation management, improves track traffic service level.
Background technology
As effective mode for alleviating urban traffic blocking, track traffic is rapidly developed in recent years.Currently, in ChinaGround city is just welcoming the upsurge of subway construction.Particularly in big city as Beijing, Shanghai, track traffic constantly improve and hairExhibition, operating line is developed into multi-thread by single line and gradually stepped into the networking stage, and urban track traffic haulage level is significantly carriedIt is high.But also attracted the huge volume of the flow of passengers simultaneously, particularly in early evening peak, section passenger flow is usually beyond the actual fortune of networkCan, so as to generate crowded bottleneck.Effective identification of the crowded bottleneck of rail network can be provided reliably for operation managementUse for reference, preferably to play the benefit of urban track traffic, adjust order of classes or grades at school of dispatching a car, the departure interval such as appropriate, multiplied with reducingVisitor's collecting and distributing time AT STATION, comfortableness when passenger takes is improved, ensured safety.Clothes can be selected for traveler travel behaviourBusiness, traveler may choose whether to avoid bottleneck interval, bottleneck time etc..It can be set for Metro Network optimization design, new lineCount based theoretical etc..Therefore, the crowded bottleneck of identification rail network is extremely necessary.
The research of traffic bottlenecks Study on Problems earliest primarily directed to road bottleneck problem, so in road bottleneck identification sideThe achievement in research in face is very abundant.In terms of microcosmic road section traffic volume bottleneck identification：It is to hand over that countries in the world, which are studied more,Lead to crowded automatic identification (Automatic Traffic congestion identification, ACI) algorithm.The ACI of early stageThe main study subject of algorithm is the traffic events of burst, the magnitude of traffic flow for being largely all based on being gathered with induction coil,The traffic data such as ground spot speed and occupation rate；Thereafter the indirect ACI algorithms developed are except the conventional Vehicle Detection instrument institute of applicationOutside the magnitude of traffic flow, occupation rate and the place speed data of offer, the basic data applied also including vehicle instantaneous velocity withJourney time, time headway, the average travel speed of traffic flow and average travel time etc.；In recent years, video image is also ACIUsed in algorithm.Have in terms of bottleneck road identification of the macroscopic view based on road network：Recognition methods based on rough set (hides dawn peak to be based onRoad network bottleneck road recognition methods [J] the highway communications science and technology of rough set, 2009,26 (9):120-124.), based on minimal cutThe dynamic traffic bottleneck identification method for collecting max-flow (is handed in Liu Jin rosy clouds urban road network traffic bottlenecks Study of recognition [D] LanzhouLogical university, 2015.), recognition methods (Li D, Fu B, Wang Y, the et al.Percolation based on seepage theorytransition in dynamical traffic network with evolving critical bottlenecks.[J].Proceedings of the National Academy of Sciences of the United States ofAmerica,2015,112(3):669-72.) etc..The abundant achievement in research of road traffic bottleneck identification is for track traffic bottleneckIdentification has certain reference, but road traffic has certain difference with track traffic, so road traffic bottleneck is knownMethod for distinguishing is not fully appropriate for track traffic bottleneck identification.
Just deployed in recent years for the research in terms of track bottleneck identification, related research is also fewer.For orbital stationThe research of point bottleneck mainly recognizes bottleneck (Zhang Qi, the Han Bao of Intra-site according to station facilities service level, queuing time etc.Bright, Li get Wei waits the Chinese railway sciences of method for evaluating operating effect of urban track traffic hub [J] of based on emulation technology,2011,32(5):120-126. Chen Feng, Wu Qibing, Zhang Huihui, wait Beijing Metro Line 1 terminal facilitieses and passenger flow relationship analysis[J] Traffic transport system engineerings and information, 2009,9 (2):93-98.).It is mainly for network bottleneck station recognitionBased on the distribution of Utopian network passenger flow, and on this basis, the station bottleneck recognition methods based on service level is built(Wang Zhipeng, military remote duckweed urban mass transit network bottleneck recognition methods [J] Chang An University journal natural science edition,2015(s1):198-202.).On the one hand, studied for the bottleneck identification of rail network, the research ten based on macro network levelDivide scarcity；On the other hand the research of existing macro network level, its Research foundation is all ideal network bus traveler assignment, with realityRail network passenger flow have a certain distance.
The content of the invention
Goal of the invention：
Focused mostly on for the existing research on track bottleneck identification in static bottleneck characteristic analysis and identification AT STATION,Lack for because passenger flow change and caused by bottleneck analysis, the bottleneck identification of network level is very related to less.The present invention is from grandA kind of angle of sight, based on reliable rail network passenger flow data, it is proposed that the crowded bottleneck of rail network based on seepage theoryRecognition methods.
Technical scheme：
A kind of crowded bottleneck identification method of rail network based on seepage theory, comprises the following steps：
1. by rail network it is abstract be an oriented topological diagram；
2. from characteristic index of the interval load factor as rail network state of section；
3. dynamic track network is built according to seepage theory, particularly：It is specific fully loaded given one of each momentRate value l, it is interval for each, if its load factor is more than particular value l, it is deleted from network, if its load factor is less thanParticular value l then retains the interval；
The change of 4.l values：For synchronization, different network states can be obtained by changing l values, and l span isThe zero interval load factor value of maximum inscribed to rail network when corresponding.During l changes to minimum value from maximum, network fromOne connected graph resolves into several small connected graphs, and the quantity and size of connected graph can all change；
5. the determination of threshold limit value：During l values are tapered into, the connected graph G for having maximum is connected with second largestScheme SG, according to seepage theory, corresponding l values are threshold limit value when SG scale reaches maximum；
6. the determination of bottleneck set：The Interval Set removed when l value changes are to threshold limit value is combined into possible bottleneck set；
7. the determination of bottleneck：Change load factor interval in possible bottleneck set one by one, if its change can to faceBoundary's threshold value changes the then interval and is defined as crowded bottleneck.
The features of the present invention：
The basis of the present invention is reliable, substantial amounts of rail network passenger flow data, rather than Utopian network passenger flow pointMatch somebody with somebody, so the achievement in research of the present invention can more accurately reflect the crowded bottleneck of reality of rail network.Bottleneck then due toThe characteristic such as concurrency, propagated, instantaneity, is that its identification brings bigger difficulty, but dynamic bottle with certain unstabilityThe identification of neck has more realistic meaning.And the present invention enters action based on seepage theory exactly from macro network aspect to rail networkThe crowded bottleneck identification of state, contributes to track operation management person to hold track running situation on the whole, so that efficiency of operation is improved,Also preferably serve passengers simultaneously.
Brief description of the drawings
Fig. 1 is the topological diagram of Beijing Rail Transit network；
Fig. 2 is situation of change of the G and SG scales with l；
Fig. 3 lists possible crowded bottleneck；
Fig. 4 differentiates to bottleneck interval；
Fig. 5 is the crowded bottleneck determined.
Embodiment
Below in conjunction with drawings and examples, the present invention is furture elucidated.The present invention provides a kind of rail based on seepage theoryRoad network congestion bottleneck identification method, methods described step is as follows：
1. by rail network it is abstract be an oriented topological diagram：
Rail network model, i.e. urban railway station are built from the P space-wises in complex network model construction method to be considered asNode, if two websites have straightforward line, then they just have even side.The rail network that this example is used is Beijing rail networkNetwork, its topological structure is as shown in Figure 1.
2. from characteristic index of the interval load factor as rail network state of section：
The computational methods of interval load factor are：
In formula, LijInterval ij load factor is represented, m represents the vehicle number by interval ij in timing statisticses, and Q represents vehicleK actual passenger number, C represents vehicle k rated passenger capacity, and the timing statisticses that this example is used are 5 minutes.
3. dynamic track network is built according to seepage theory, particularly：It is specific fully loaded given one of each momentRate value l, it is interval for each, if its load factor is more than particular value l, it is deleted from network, if its load factor is less thanParticular value l then retains the interval.Have
The change of 4.l values：For synchronization, different network states can be obtained by changing l values, and l span isThe zero interval load factor value of maximum inscribed to rail network when corresponding.During l changes to minimum value from maximum, network fromOne connected graph resolves into several small connected graphs, and the quantity and size of connected graph can all change.That is, according to step 3,Change l values, the adjacency matrix of different rail networks can be obtained.
5. threshold limit value lcDetermination：During l values are tapered into, it may appear that maximum connected graph G and second largestConnected graph SG, according to seepage theory, when SG scale reaches maximum, corresponding l values are threshold limit value lc.Fig. 2 is G and SGScale is with l situation of change, and data time is 12 days 08 May in 2016:20—08:25, when can determine that the statistics by Fig. 2Interior lcFor 0.6.
6. the determination of bottleneck set：The Interval Set removed when l value changes are to threshold limit value is combined into possible bottleneck set.By the l knowable to step 5 nowcFor 0.6, it is hereby achieved that possible bottleneck set.Fig. 3 is now to be moved under threshold limit valueThe rail network removed is interval, and one has 3 intervals, is represented in figure for dotted line, in order to become apparent from having roundlet on three intervalsCircle is indicated.
7. the determination of bottleneck：Change load factor interval in possible bottleneck set one by one, if its change can to faceBoundary's threshold value changes the then interval and is defined as crowded bottleneck.
On the basis of step 6,3 interval load factors are reduced respectively, and the threshold limit value of network may change, it is also possible toDo not change.It can determine that by Fig. 4, the change of interval 1 load factor improves network state, and interval 2 and interval 3 do not have then, soInterval 1 is the crowded bottleneck of rail network now, its position such as Fig. 5.