CN103218670B - Urban railway traffic random passenger flow loading method - Google Patents

Abstract

The invention discloses an urban railway traffic random passenger flow loading method which comprises the following steps: 1.1, initializing, dividing one day into a plurality of time windows averagely, loading basic data, at the same time, setting the total number of the time windows as n, setting and marking an initial value of a variable t of the time windows as 1, and setting an initial value CN1 of road congestion charge as 0; 1.2, reading an origin and destination (OD) distribution volume of a tth time window, based on a basically effective path collection, calculating time of a first regular bus and a last regular bus of each path, judging the relationship of the time and the tth time window, and generating a dynamic effective path collection; 1.3, based on the dynamic effective path collection and an urban railway traffic passenger going-out path selection model, confirming a selected portion of each effective path between the OD, and obtaining a path flow; 1.4, performing statistics on a passenger flow of each section of a road, and updating the road congestion charge CNt+1; and 1.5, stopping the judgment, if the t is less than n, t=t+1 and continuing to iterate from the step 1.2, or, stopping the method.

Description

A kind of urban railway traffic random passenger flow loading method

Technical field

The present invention relates to technical field of rail traffic, more particularly, to a kind of urban railway traffic random passenger flow loading method.

Background technology

It is the final link of urban track traffic for passenger flow requirement forecasting that urban mass transit network random passenger flow loads, and passes through Passenger flow loads and not only can obtain the important passenger flow index such as section flow, transfer amount and circuit flow to instruct train capacity to be equipped with Work out moreover it is possible to obtain the freight allocating table of different subjects of operation with train operation plan, the passenger flow of therefore efficiently and accurately is loaded with Important realistic meaning.

Current urban mass transit network random passenger flow loading method, is divided into the loading theoretical based on collection meter and based on non- The loading of collection meter theoretical (Logit or Probit model).Macroscopic view is shown that based on the theoretical random passenger flow loading result of collection meter Statistical law, and the random passenger flow based on non-maximal suppression loads the microcosmic choosing that subordinate act Angle of Interpretation analyzes passenger's trip route Select behavior, thus the random passenger flow based on non-maximal suppression load more can reflect reality, more accurate.But it is currently based on non-collection meter reason By urban mass transit network random passenger flow load, lack and consider that road network topology structure and path are moved towards to passenger's Path selection Impact；General piecewise function represents Congestion surcharge with it is impossible to reflect the company that Congestion surcharge affects on passenger's Path selection preference Continuous change, and partial parameters therein are typically derived from empirical value；General first and last regular bus of ignoring limits to passenger flow loading result Impact.

Content of the invention

In order to overcome the shortcomings of prior art construction, the present invention provides a kind of urban railway traffic random passenger flow loading side Method.

The embodiment of the invention discloses a kind of urban railway traffic random passenger flow loading method, comprise the following steps：

1.1st, initialize, one day is averagely divided into multiple time windows, and load road network topology data, time-table, OD abundance under each time window, concurrently setting time window sum is n, the initial value of variable t setting labelling time window as 1, if Determine the initial value CN of the crowded fee forecast in section₁=0；

1.2nd, read the OD abundance of t time window, based on basic active path set, calculate the first and last regular bus in each path Time, judge the relation of this time and t time window, generate dynamic effective paths set RSET_t；

1.3rd, it is based on dynamic effective paths set RSET_tWith urban track traffic passenger's route choice model, determine Each active path selection percentage between OD, obtains path flow, wherein, the effect of urban track traffic passenger's route choice model The urban track traffic passenger's travel route choice effect set up when being built with active path set basic in function synchronizing rapid 1.2 Use function；

1.4th, count the volume of the flow of passengers in each section, update section Congestion surcharge CN_t+1；

1.5th, terminate judging, if t<N, then t=t+1, continue iteration from step 1.2；Otherwise, the method terminates.

In described step 1.2, the structure of basic active path set comprises the following steps：

2.1st, build urban track traffic passenger's travel route choice utility function, the variable in utility function is except ticket Valency, time out of the station, riding time, platform Waiting time, transfer travel time, number of transfer and crowded cost element, also draw Enter angle expense to express the impact to passenger's Path selection of road network topology and path trend；

2.2nd, it is based on equivalent riding time coefficient, other factors are all converted into riding time；

2.3rd, determine the tolerance threshold to the time for the passenger；

2.4th, determine basic active path set using Double-sweep searching algorithm.

The present invention carries out adding first and last regular bus time restriction condition when passenger flow loads, and constructs dynamic effective paths collection Close, and kernel model introduces angle expense and expresses the impact to passenger's Path selection of road network structure and path trend, will be crowded Expense is processed as the product of compartment load factor and section riding time, with Same Function reflection Congestion surcharge to passenger's Path selection The consecutive variations of preference impact, improve loading reasonability and path flow estimation precision.

Brief description

When considered in conjunction with the accompanying drawings, by referring to detailed description below, can more completely more fully understand the present invention with And easily learn the adjoint advantage of many of which, but accompanying drawing described herein is used for providing a further understanding of the present invention, Constitute the part of the present invention, the schematic description and description of the present invention is used for explaining the present invention, does not constitute to this Bright improper restriction, wherein：

Fig. 1 is embodiment of the present invention urban railway traffic random passenger flow loading method flow chart.

Fig. 2 is angle expense exemplary plot.

Specific embodiment

With reference to Fig. 1-2, embodiments of the invention are illustrated.

Understandable for enabling above-mentioned purpose, feature and advantage to become apparent from, right with reference to the accompanying drawings and detailed description The present invention is described in further detail.

As shown in figure 1, a kind of urban railway traffic random passenger flow loading method, the specific embodiment of the method is as follows.

S1, initialization：

(1) averagely it is divided within one day each time window, the 1st time window is daily initial time section, when setting labelling Between window variable t initial value be 1, time window sum be n.

(2) road network topology data, time-table, the basic data such as OD abundance under each time window are loaded.

(3) initial value of section Congestion surcharge is 0, i.e. CN₁=0.

S2, OD (starting point to the end) abundance of reading t time window, based on basic active path set, calculate each road The first final vehicle hour in footpath, judge the relation of this time and t time window, generate dynamic effective paths set RSET_t(during t Between window active path set).

Dynamic effective paths set RSET_tGeneration method, be will be each in basic active path set under t time window Following process is done in path：The final vehicle hour of end time or path that the first vehicle hour in path is more than t time window is less than The path of the initial time of t time window does not consider；The first vehicle hour in path is in path between t time window by normal Pass considers；The path that the final vehicle hour in path is between t time window with the final vehicle hour for Detailizing joint is Little time window, and the proportion of time window according to shared by little time window determines OD abundance under little time window, this path is in this hour Between can be regarded as normal pass path under window；All normal pass paths are active path, and the active path under t time window The set of composition is dynamic effective paths set RSET_t.

S3, be based on RSET_t(time window is in morning peak, evening height with urban track traffic passenger's route choice model Peak or flat peak phase, model parameter is different), determine each active path selection percentage between OD, obtain path flow.

${\mathrm{ql}}_k^{rs}=\frac{\exp \left(V_k^{rs}\right)}{\mathrm{\&Sigma;}\exp \left(V_k^{rs}\right)}\&times;q_{rs}$

In formula,It is the path flow of the kth bar active path to the OD of terminal s from starting point r, q_rsIt is from r to s Timesharing OD abundance, unit be person-time；It is the Path selection effectiveness of kth bar active path between the OD from r to s, by S2 Basic active path set construction when urban track traffic passenger's travel route choice utility function of being set up determine.

S4, the section based on path flow, path are constituted and time-table, update section Congestion surcharge CN_t+1, section Congestion surcharge is the product of section load factor and section riding time, that is,

CN_t+1=qplink_t*TRlink (2)

Wherein, CN_t+1It is that the section Congestion surcharge updating in t+1 time window is used, unit is hour；qplink_tFor The section load factor of t time window；TRlink is the riding time in section, and unit is hour.

S5, termination judge.If t<N, then t=t+1, continue iteration from S2；Otherwise, the method terminates.

S6, path flow and the link flow exporting under each time window.

Wherein, in above-mentioned S2, the building process of basic active path set is as follows

(1) build urban track traffic passenger's travel route choice utility function.

Based on MNL (Multinomial Logit) model, construct urban track traffic passenger's travel route choice effectiveness Function, this function not only allows for admission fee, time out of the station, riding time, platform Waiting time, transfer travel time, transfer The factor such as number of times and Congestion surcharge, also introduces angle expense to express road network topology and path trend to passenger's Path selection Congestion surcharge is processed as the product of compartment load factor and respective stretch run time by impact simultaneously.

$V_k^{rs}=-{\mathrm{\&Sigma;}}_{i=1}^m{\mathrm{\&beta;}}_i{X}_{ik}^{rs}---\left(3\right)$

In formula,It is the determination item of the Path selection effectiveness of kth paths between the OD from r to s；It is from r to s Between OD, the ith feature property value of kth paths is (when including the cars such as riding time, admission fee, travel time out of the station, platform Between, transfer travel time, transfer Waiting time, number of transfer, Congestion surcharge with and the factor such as angle expense (path trend))；β_i Parameter for individual features attribute；M is influence factor's number.

Wherein, angle expense, such as Fig. 2, are that move towards deviation straight from origin-to-destination in each section to passenger's trip route Reach the punishment in direction, this value should be incremented by with the increase of deviation angle and road section length, and variation tendency is also with deviation angle It is incremented by and increases, that is, angle cost function is that single order can be led, and derived function is greater than 0 increasing function, is expressed as follows：

The computing formula of described angle expense is as follows：

${\mathrm{AC}}_k^{rs}=\underset{p=1}{\overset{N}{\&Sigma;}}{L}_p*tan({\mathrm{\&theta;}}_p/4)---\left(4\right)$

In formula：It is the angle expense of kth paths between the OD from r to s；L_pPth bar section for this paths Length；θ_pAngle from the through direction of origin-to-destination is deviateed in pth section for this paths, and value is [0, π], and N is The section number of this paths；

Urban track traffic passenger's travel route choice utility function considers admission fee, time out of the station, riding time, stands The factors such as platform Waiting time, transfer travel time, number of transfer, Congestion surcharge use and angle expense (path trend).Especially angle The introducing of degree expense correct for the ignorance of the road network topology and path trend impact to passenger's Path selection all the time.

A. admission fee.Urban track traffic charging has two ways, and one is " flat fare "；Another kind is to divide by traveling mileage Section charging, thus admission fee p is expressed as p=δ * p'.In formula, δ is mark amount, is 0 when " flat fare ", is 1 during pricing for segment；p' It is admission fee constant when " flat fare ", be admission fee piecewise function during pricing for segment.

B. the time out of the station, mainly include from enter the station time and the going out from platform to outbound of swiping the card into CFS to CFS platform of swiping the card Stand the time, represented with IOt.

C. riding time, it is mainly included in the run time in compartment and bus stoppage time, is represented with IVt.

D. platform Waiting time, refers to the time waiting vehicle to arrive on platform, is represented with Wt.Waiting time is main Affected in the bus stoppage time (TSt) of this platform and compartment load factor (qplink) by departure interval (Int), this car it is assumed that The arrival of passenger is obeyed and is uniformly distributed, and passenger always once can get on the bus when vehicle reaches twice, then platform Waiting time is： Wt=(Int+TSt)/2+ δ '^*Int.Wherein, work as qplink>When 120%, δ ' takes 1, otherwise takes 0.

E. change to travel time, refer to from a train to the travel time of another train, represented with Trt.

F. number of transfer, is represented with Trs.

G. Congestion surcharge is used, and represents that path Congestion surcharge is used with VP.

$VP=\frac{\underset{u=1}{\overset{l}{\&Sigma;}}({\mathrm{CN}}_{u,t-1}*{\mathrm{\&delta;}}_{k,u}^{rs})}{\underset{u=1}{\overset{l}{\&Sigma;}}\left({\mathrm{\&delta;}}_{k,u}^{rs}\right)}---\left(5\right)$

Wherein, VP uses for path Congestion surcharge；CN_u,t-1It is the section updating when the t-1 time window (i.e. t-1 iteration) Section Congestion surcharge (the unit of u：Hour)；If on section u kth paths between the OD from r to s,For 1, otherwise for 0.

H. angle expense (Angular Cost), is represented with AC, is tried to achieve using formula (4).

Passenger is different in the travel route choice preference of morning peak, evening peak peace peak phase, using Maximum Likelihood Estimation Method, T value method of inspection and goodness of fit criterion carry out parameter calibration, and result is as follows：

${\mathrm{MV}}_k^{rs}=-0.12{\mathrm{IOt}}_k^{rs}-2.9{\mathrm{IVt}}_k^{rs}-0.21{\mathrm{Wt}}_k^{rs}-0.24{\mathrm{Trt}}_k^{rs}-3.2{\mathrm{Trs}}_k^{rs}-5.8{\mathrm{VP}}_k^{rs}-4.03{\mathrm{AC}}_k^{rs}---\left(6\right)$

${\mathrm{EV}}_k^{rs}=-2.2{\mathrm{IOt}}_k^{rs}-3.9{\mathrm{IVt}}_k^{rs}-1.36{\mathrm{Wt}}_k^{rs}-0.58{\mathrm{Trt}}_k^{rs}-4.8{\mathrm{Trs}}_k^{rs}-7.22{\mathrm{VP}}_k^{rs}-5.1{\mathrm{AC}}_k^{rs}---\left(7\right)$

${\mathrm{CV}}_k^{rs}=-0.19{\mathrm{IOt}}_k^{rs}-2.91{\mathrm{IVt}}_k^{rs}-0.1{\mathrm{Wt}}_k^{rs}-0.52{\mathrm{Trt}}_k^{rs}-3.9{\mathrm{Trs}}_k^{rs}-54.7{\mathrm{VP}}_k^{rs}-3.03{\mathrm{AC}}_k^{rs}---\left(8\right)$

In formula：For during morning peak between the OD from r to s kth paths Path selection effectiveness；High for evening During peak between the OD from r to s kth paths Path selection effectiveness；For kth paths between the OD from r to s during flat peak Path selection effectiveness.Because the Data Source of the present invention is in " a ticket system " urban mass transit network, therefore there is no admission fee factor, I.e. δ=0.

(2) equivalent time conversion, based on formula (6) (7) (8), is utilized respectively formula (9), you can obtain morning peak, evening height The equivalent riding time of peak peace peak phase each factor.

ETTC_i=β_i/β₁(9)

In formula, ETTC_iEquivalent riding time coefficient for ith attribute；β_iFor ith attribute coefficient in a model；β₁For taking advantage of Car time coefficient in a model.

(3) passenger to time degrees of tolerance analyze, passenger the tolerance of time is included the absolute tolerance to the time with relative Tolerance, that is,

T_i≤ min (tt+H, C*t) (10)

In formula, T_iTotal time for certain paths between OD pair；Tt is the total time of the shortest path between OD pair；H is absolute threshold Value；C is relative ratio.Suggestion H=13.39 minute, C=2.7.

(4) utilize Double-sweep algorithm, filter out the every one K bar shortest path between OD pair, and then pass through passenger couple The degrees of tolerance of time limits, and deletes the path of the condition that is unsatisfactory for, constructs basic active path set.

Although the foregoing describing the specific embodiment of the present invention, it will be appreciated by those of skill in the art that these Specific embodiment is merely illustrative of, those skilled in the art in the case of the principle without departing from the present invention and essence, Various omissions, substitutions and changes can be carried out to the details of said method and system.For example, merge said method step, thus Then belong to the scope of the present invention according to the substantially identical function of substantially identical method execution to realize substantially identical result.Cause This, the scope of the present invention is only limited by the claims that follow.

Claims (1)

1. a kind of urban railway traffic random passenger flow loading method is it is characterised in that comprise the following steps：

1.1st, initialize, one day is averagely divided into multiple time windows, and load road network topology data, time-table, each when Between OD abundance under window, concurrently setting time window sum is n, the initial value of variable t setting labelling time window as 1, setting road The initial value CN of section Congestion surcharge₁=0；

1.2nd, read the OD abundance of t time window, based on basic active path set, during the first and last regular bus in each path of calculating Between, judge the relation of this time and t time window, generate dynamic effective paths set RSET_t；Wherein, basic valid path set The building process closing is shown in step 2.1-2.4, and dynamic effective paths set RSET under t time window_tGenerating process as follows：Road The first vehicle hour in footpath be more than the end time of t time window or path final vehicle hour be less than t time window initial when The path carved does not consider；The path that the first vehicle hour in path is between t time window is considered by normal pass path；Path Final vehicle hour be in path between t time window with the final vehicle hour for Detailizing joint for little time window, and according to little The proportion of time window shared by time window determines OD abundance under little time window, and normal pass be can be regarded as under this little time window in this path Path；All normal pass paths are active path, and the set of the active path composition under t time window is dynamically Active path set RSET_t；

1.3rd, it is based on dynamic effective paths set RSET_tWith urban track traffic passenger's route choice model, calculate t when Between allocation proportion on each active path for the OD abundance under window, the OD abundance under t time window is multiplied by this ratio, that is, obtains Obtain the path flow in each path under t time window；

The computing formula of described path flow is as follows：

${\mathrm{ql}}_k^{rs}=\frac{\exp \left(V_k^{rs}\right)}{\&Sigma;\exp \left(V_k^{rs}\right)}\&times;q_{rs}$

In formula,It is the path flow of the kth bar active path to the OD of terminal s from starting point r, q_rsIt is the timesharing from r to s OD abundance,It is the determination item of the Path selection effectiveness of kth bar active path between the OD from r to s, this determination item is synchronous Urban track traffic passenger's travel route choice utility function that in rapid 1.2, basic active path set is set up when building is really Determine item；

1.4th, the section based on path flow, path is constituted and time-table, updates section Congestion surcharge CN_t+1, section is crowded Expense is the product of section load factor and section riding time, i.e. CN_t+1=qplink_t* TRlink, this value will be used as next time Calculate the input data of path flow in step 1.3 during iteration；Wherein, CN_t+1It is the road updating in t+1 time window Section Congestion surcharge is used, and unit is hour；qplink_tSection load factor for t time window；TRlink is the riding time in section, Unit is hour；

1.5th, terminate judging, if t<N, then t=t+1, continue iteration from step 1.2；Otherwise, the method terminates；

In described step 1.2, the structure of basic active path set comprises the following steps：

2.1st, build urban track traffic passenger's travel route choice utility function, the computing formula of the determination item of this function is such as Under：

$V_k^{rs}=-{\&Sigma;}_{i=1}^m{\mathrm{\&beta;}}_i{X}_{ik}^{rs}$

In formula,It is the determination item of the Path selection effectiveness of kth paths between the OD from r to s；Be OD from r to s it Between kth paths ith feature property value；β_iParameter for individual features attribute；M is characterized the number of attribute i；

Above-mentioned characteristic attribute value not only comprise describe urban track traffic service level variable, that is, the riding time in path, Admission fee, travel time out of the station, platform Waiting time, transfer travel time, transfer Waiting time, number of transfer and Congestion surcharge With also comprising to describe the variable that the non-rectilinear degree of road network topology and trip route affects on passenger's Path selection, that is, angle is taken With variable, wherein, the value of path Congestion surcharge is the section Congestion surcharge CN that this path respectively forms section_tSum is divided by section Bar number；Angle expense is that each section is moved towards to deviate the punishment from the through direction of origin-to-destination, institute to passenger's trip route The computing formula stating angle expense is as follows：

${\mathrm{AC}}_k^{rs}=\underset{p=1}{\overset{N}{\&Sigma;}}{L}_p*tan({\mathrm{\&theta;}}_p/4)$

In formula：It is the angle expense of kth paths between the OD from r to s；L_pLength for the pth bar section of this paths Degree；θ_pAngle from the through direction of origin-to-destination is deviateed in pth bar section for this paths, and N is the section of this paths Bar number；

2.2nd, it is based on equivalent riding time coefficient, other factors are all converted into equivalent riding time；

2.3rd, determine the tolerance threshold to the time for the passenger；

2.4th, determine basic active path set using Double-sweep searching algorithm.