CN106327871A - Highway congestion forecasting method based on historical data and reservation data - Google Patents

Abstract

The invention discloses a highway congestion forecasting method based on historical data and reservation data. The method comprises steps: 1) a researched highway is selected and basic data are acquired, wherein the basic data altogether comprise data in four aspects: highway user reservation data, highway road network basic data, highway traffic flow historical data and highway toll policy data; 2) according to the road network basic data and the user reservation data, the highway road network is subjected to spatio-temporal partition processing, and a forecasting time period for each road section is determined; 3) the traffic amount of each road section in the highway road network in the forecasting time period is forecasted; 4) the congestion degree of each road section is forecasted; and 5) a programming language is used for realizing visual output of the congestion forecasting result. The problems that the traditional congestion forecasting method can not integrate the reservation data and can not reflect changes of the distribution law of the traffic amount brought by reservation passing in the highway road network can be solved, and more reliable basis is provided for trip guidance and highway management.

Description

A kind of crowded Forecasting Methodology of highway merging historical data and reservation data

Technical field

The present invention relates to the technical field of the crowded prediction of highway, refer in particular to a kind of fusion historical data and reservation number According to the crowded Forecasting Methodology of highway.

Background technology

Along with being continuously increased of volume of transport, the congestion problems (especially in peak period festivals or holidays) of China's highway The most serious.The first day on National Day in 2012, the whole nation had 26 highways in 16 provinces to occur blocking up；And National Day in 2015 The first day, only Guangdong Province just have 31 highways to occur blocking up.In order to solve congestion problems, freeway management department is often Block up and take after occurring certain to dredge measure.Although these measures can reach certain effect, but blocking up of having occurred Loss is all caused to traveler and manager.Therefore, some scholars propose and carry out setting of congestion dispersal before generation of blocking up Think.The most ripe development of Mobile Internet technology makes this imagination be possibly realized: by mobile Internet and highway Operation management combines, and the active that current by reservation in induced travel realizes blocking up is dredged.

Realizing such target, highway is carried out the most crowded prediction is premise.Chinese scholars is the most right The crowded prediction of highway has carried out substantial amounts of research.Developed country starts from the middle of last century to the research of crowded forecast model. China's research in terms of road traffic congestion prediction start late and great majority research be all by the short-term volume of traffic or other Index carrys out evaluation path Traffic Congestion.The main method in terms of crowded prediction is applied to have time series models, card at present Kalman Filtering model, neural network model, supporting vector machine model etc..These methods have his own strong points, but have one jointly Point is predicted on the basis of historical data.This reservation making existing method be difficult to tackle highway is passed through as band The challenge come: 1. reservation data is by the data source new for the offer of crowded prediction work, reservation number to be considered during prediction According to the fusion with historical data.2. reservation behavior will change the volume of traffic distribution at freeway network.Predicted if crowded These changes are not accounted for by journey and applies mechanically existing method simply, it will cause the decline of crowded precision of prediction.

Therefore, the present invention proposes a kind of fusion reservation data and historical data carries out the crowded Forecasting Methodology of highway, More reliable foundation is provided for induced travel and freeway management work.

Summary of the invention

It is an object of the invention to overcome the shortcoming and defect of prior art, it is provided that a kind of fusion historical data and reservation number According to the crowded Forecasting Methodology of highway, break through the conventional crowded Forecasting Methodology of highway do not consider user's reservation data, cannot Reflection carry out reservation current after volume of traffic defect such as redistribution in road network, it is achieved that under the conditions of highway reservation is current Crowded prediction, it is possible to being effectively improved precision of prediction, the realization current for highway reservation lays the foundation.

For achieving the above object, technical scheme provided by the present invention is: a kind of historical data and reservation data of merging The crowded Forecasting Methodology of highway, comprises the following steps:

1) highway of selected research, obtains basic data, altogether includes the data of four aspects: highway user is pre- About data, freeway network basic data, freeway traffic flow historical data, expressway tol lcollection policy data.

2) according to highway user reservation data and freeway network basic data, when freeway network is carried out Empty division processes and determines the prediction period in each section.

3) volume of traffic in section each to freeway network in prediction period is predicted: first, utilize historical data pair Traffic total amount is predicted；Then, total wheel traffic is allocated and by reservation data to non-according to reservation amount and non-reservation amount Reservation amount is modified；Finally reservation amount after distribution and non-reservation amount are added the volume of traffic obtaining each section.

4) crowding in each section is predicted: employing saturation is as crowded evaluation index, by each section of threshold decision Congestion state.

5) programming language is utilized to realize the crowded visualization predicted the outcome output.

In step 1) in, described highway user reservation data includes departure place, destination that user inputs, when entering the station Section, vehicle, user app or the market share of wechat public number, obtained by user app backstage or wechat public number backstage Take.The road network basic data of described highway include the length of each road, number of track-lines, lane width, design capacity, from By the time of vehicle operation under the conditions of stream, by Operation and Management of Expressway, department obtains.Described freeway traffic flow history Data include flow and vehicle speed data, and by Operation and Management of Expressway, department obtains.Described expressway tol lcollection policy data Being worth including basic rate, the margin of preference and hourage, by Operation and Management of Expressway, department obtains.

In step 2) in, according to freeway network basic data and user's reservation data, freeway network is carried out Space-time division processes and determines the prediction period in each section, specifically includes following steps:

2.1) division of time: when using the unit consistent with the Time segments division that enters the station on user app or wechat public number Between interval of delta t, can be according to 10min, 20min, 30min, the 60min to be asked for of Operation and Management of Expressway department, one day 24 Hour it is divided into m timeslice.

$\left\{\begin{array}{c}T=\{t_1,...,t_i,...,t_m\}\\ m=\frac{1440}{\mathrm{\Δ}t}\\ i=1,...,m\end{array}\right.$

In formula: T timeslice set；

t_iT_iIndividual timeslice；

M timeslice number, individual；

Δ t unit interval, min.

2.2) division in space: use and the pavement section phase one of Operation and Management of Expressway department collection traffic flow data Unit space interval delta s caused, can be according to the 1km to be asked for of Operation and Management of Expressway department, and 2km, 5km, 10km etc., by road All highways in net are divided intoIndividual section.

$\left\{\begin{array}{c}S=\{s_{11},s_{12},...,s_{1k_1},...,s_{j1},s_{j2},...,s_{{\mathrm{jk}}_j},...,s_{n1},s_{n2},...,s_{{\mathrm{nk}}_n}\}\\ k_j=\[\frac{L_j}{\mathrm{\Δ}s}\]\\ j=1,...,n\end{array}\right.$

In formula: S section is gathered；

The kth of j-th strip highway_jIndividual section；

k_jThe section number of j-th strip highway, individual；

L_jThe length of j-th strip highway, km；

Δ s unit space is spaced, km；

N highway bar number, bar.

2.3) prediction period in each section is determined: user is arrived the time of each road segment end and calculated by equation below:

$\left\{\begin{array}{c}{t}_{s_{{\mathrm{jk}}_j}}=t_{s_{j(k_j-1)}}+\frac{\mathrm{\Δ}s}{v_j}\\ t_{s_{j1}}=\frac{t_a+t_b}{2}\end{array}\right.$

In formula:User reaches the kth of j-th strip highway_jThe time of individual road segment end；

User reaches the time of the entrance of j-th strip highway；

v_jThe average speed of operation of j-th strip highway；

t_aThe starting point of the period of entering the station of user's reservation；

t_bThe terminal of the period of entering the station of user's reservation.

In this manner it is possible to judgement sectionPrediction period: ifThen sectionPrediction period be t_i。

In step 3) in, the volume of traffic in section each to prediction period freeway network is predicted, and specifically includes following Step:

3.1) predict the total wheel traffic in each section according to historical data, use BP neutral net to be predicted, can directly exist Matlab calls the function carried realize.

3.2) reservation traffic total amount is calculatedWith non-reservation traffic total amountIts computing formula is:

$\left\{\begin{array}{c}{Q}_{t_i}=\underset{j}{\Σ}\underset{k_j}{\Σ}{q}_{{\mathrm{jk}}_j}^{t_i}\\ Q_{t_i,reservation}=Q_{t_i}\·\mathrm{\α}\\ Q_{t_i,non-reseration}=Q_{t_i}-Q_{t_i,reservation}\end{array}\right.$

In formula:Prediction period t_iTime road network total wheel traffic；

Prediction period t_iTime road network reservation traffic total amount；

Prediction period t_iTime road network non-reservation traffic total amount；

Prediction period t_iTime j-th strip highway kth_jThe volume of traffic in individual section；

α user app or the market share of wechat public number.

Remaining each parameter meaning is as previously mentioned.

3.3) according to the historical rethinking rule of vehicle flowrate, non-reservation traffic total amount is allocated in the range of road network, To the non-reservation amount in each section, its computing formula is:

$q_{{\mathrm{jk}}_j,non-reservation}^{t_i}=\frac{q_{{\mathrm{jk}}_j}^{t_i}}{Q_{t_i}}\·Q_{t_i,non-reservation}$

In formula:Prediction period t_iTime j-th strip highway kth_jThe non-reservation traffic in individual section Amount.

3.4) use user equilibrium model that reservation traffic total amount is allocated in the range of road network, obtain the pre-of each section About measure.

$\min Z\left(X\right)=\underset{j}{\mathrm{\Σ}}{\∫}_0^{q_{{\mathrm{jk}}_j,non-reservation}^{t_i}}{t}_{q_{{\mathrm{jk}}_j,non-reservation}^{t_i}}\left(w\right)dw$

$S.t.\left\{\begin{array}{c}\underset{j}{\Σ}{f}_{t_{i}j}=q_{{\mathrm{jk}}_j,non-reservation}^{t_i},\\ f_{t_{i}j}\≥0,\end{array}\right.$

$q_{{\mathrm{jk}}_j,non-reservation}^{t_i}=\underset{j}{\Σ}{f}_{t_{i}j},\∀j$

In formula:Prediction period t_iTime j-th strip highway flow.

Subscriber Impedance Function in road network, its concrete form is:

$t_{q_{{\mathrm{jk}}_j,non-reservation}^{t_i}}\left(w\right)=t_o\[1+\mathrm{\α}{\left(\frac{q_{{\mathrm{jk}}_j,non-reservation}^{t_i}+q_{{\mathrm{jk}}_j,reservation}^{t_i}}{C_{{\mathrm{jk}}_j}}\right)}^{\mathrm{\β}}\]+\frac{c_o(1-\mathrm{\γ})}{\mathrm{\τ}}$

In formula: t_oRunning time under the conditions of freely flowing, min；

The kth of j-th strip highway_jThe road design traffic capacity in individual section, pcu/h；

c_oBasis pass cost use, unit；

The γ margin of preference, %；

The parameter that α, β are to be calibrated；

τ is worth hourage, unit/h.

3.5) using actual reservation amount to be modified the reservation amount in each section, modification rule is as follows:

If 1.ThenConstant.

If 2.Then

In formula:Prediction period t_iTime j-th strip highway kth_jThe actual reservation in individual section is handed over Logical total amount；

3.6) each section reservation amount and non-reservation amount are overlapped, obtain the volume of traffic in each section.

$q_{{\mathrm{jk}}_j}^{t_i}=q_{{\mathrm{jk}}_j,reservation}^{t_i}+q_{{\mathrm{jk}}_j,non-reservation}^{t_i}$

In formula:Prediction period t_iTime j-th strip highway kth_jThe reservation volume of traffic in individual section.

In step 4) in, it was predicted that the crowding in each section, specifically include following steps:

4.1) calculating the saturation in each section, computing formula is:

$x_{{\mathrm{jk}}_j}^{t_i}=\frac{q_{{\mathrm{jk}}_j}^{t_i}}{C_{{\mathrm{jk}}_j}}$

In formula:Prediction period t_iTime j-th strip highway kth_jThe saturation in individual section.

4.2) division of crowding grade, is divided into different grades by saturation threshold value by crowding.Can basis The traffic survey data in each city determines threshold value.In the case of without survey data, it is proposed that use following form to carry out classification.

Table 1

In step 5) in, utilize programming language to realize the crowded visualization predicted the outcome output.By different crowded Degree grade is corresponding from different colors, gives each section when writing code by corresponding color code.Operation code, can be defeated Go out crowded predicting the outcome.

The present invention compared with prior art, has the advantage that and beneficial effect:

1, development of Mobile Internet technology is applied in freeway management, obtain by mobile phone app or wechat public number Reservation data, in order to realize, the reservation of highway is current provides basis.

2, user's reservation data is fused in prediction algorithm, the result utilizing historical data to carry out crowded prediction is carried out Revise, it is possible to increase the precision of prediction work.

3, algorithm is easier, and the process of modeling is simple, need not substantial amounts of learning sample during training, it was predicted that essence Spend higher.

4, it is easy to Operation and Management of Expressway department and carries out traffic dispersion by lever of price, utilization reservation trip mode, Promote the user equilibrium of highway network, avoid the formation of as far as possible and block up, also allow for user and avoid crowded section of highway and block up the period Staggered shifts.

5, freeway management department is conducive to grasp road grid traffic stream information and jam situation in time, to following possible shape Blocking up of becoming carries out early warning, and that takes necessity in time discongests measure.

Accompanying drawing explanation

Fig. 1 is the logic flow schematic diagram of the present invention.

Fig. 2 is the road network schematic diagram of the embodiment of the present invention.

Fig. 3 is the crowded prediction visualization output result of the embodiment of the present invention.

Detailed description of the invention

Below in conjunction with specific embodiment, the present invention is described further.

As it is shown in figure 1, the fusion historical data described in the present embodiment and the crowded Forecasting Methodology of highway of reservation data, Comprise the following steps:

1) highway of selected research, obtains basic data, altogether includes the data of four aspects: highway user is pre- About data, freeway network basic data, freeway traffic flow historical data, highway reservation cost data.Wherein, Described highway user reservation data includes departure place, the destination that user inputs, period of entering the station, vehicle, user app or micro- The market share of letter public number, is obtained by user app backstage or wechat public number backstage.The road network of described highway Basic data include the length of each road, number of track-lines, lane width, design capacity, freely flow under the conditions of vehicle travel Time, by Operation and Management of Expressway, department obtains.Described freeway traffic flow historical data includes flow and speed number According to, by Operation and Management of Expressway, department obtains.Described expressway tol lcollection policy data includes basic rate, the margin of preference Being worth with hourage, by Operation and Management of Expressway, department obtains.

The freeway network that the present embodiment is used is as shown in Figure 2.Assuming that obtain certain user input from user app backstage Departure place be A, destination be B, the period of entering the station be 8:00 9:00, vehicle is a class car；The market share of user app is 12%.

According to certain user above-mentioned reservation departure place, destination and enter the station the period, collect remaining data as follows: at a high speed Highway network basic data is as shown in table 1, the flow histories data of highway 1,2 and current subscriber total amount be respectively such as table 2, (owing to length is limited, only list the historical data of prediction period), expressway tol lcollection policy data such as table 4 shown in table 3 Shown in, the average speed of operation of two highways is 80km/h, and the time of vehicle operation under the conditions of freely flowing is respectively 1.1h And 1.5h.

Table 1 freeway network basic data

The flow histories data of table 2 highway 1 and current subscriber total amount

	8/1	8/8	8/15	8/22	8/29	9/5	9/12	9/19	Reservation
										8:00 9:00	4513	4351	4315	4415	4483	4335	4214	4377	29
9:00 10:00	4296	4021	4141	4113	4291	4061	4109	4121	14

The flow histories data of table 3 highway 2

	8/1	8/8	8/15	8/22	8/29	9/5	9/12	9/19	Reservation
										8:00 9:00	1193	1210	1189	1198	1170	1175	1161	1191	686
9:00 10:00	1121	1144	1123	1174	1098	1124	1113	1113	570

Table 4 expressway tol lcollection policy data

2) according to freeway network basic data and user's reservation data, freeway network is carried out at space-time division Manage and determine the prediction period in each section, specifically include following steps:

2.1) division of time: when using the unit consistent with the Time segments division that enters the station on user app or wechat public number Between interval of delta t, can be according to 10min, 20min, 30min, the 60min to be asked for of Operation and Management of Expressway department, one day 24 Hour it is divided into m timeslice.

$\left\{\begin{array}{c}T=\{t_1,...,t_i,...,t_m\}\\ m=\frac{1440}{\mathrm{\Δ}t}\\ i=1,...,m\end{array}\right.$

In formula: T timeslice set；

t_iT_iIndividual timeslice；

M timeslice number, individual；

Δ t unit interval, min.

In the present embodiment, unit interval Δ t=60min, within one day 24 hours, it is divided into m=24 timeslice, then T= {t₁,···,t₁₂,···,t₂₄}。

2.2) division in space: use and the pavement section phase one of Operation and Management of Expressway department collection traffic flow data Unit space interval delta s caused, can be according to the 1km to be asked for of Operation and Management of Expressway department, and 2km, 5km, 10km etc., by road All highways in net are divided intoIndividual section.

$\left\{\begin{array}{c}S=\{s_{11},s_{12},...,s_{1k_1},...,s_{j1},s_{j2},...,s_{{\mathrm{jk}}_j},...,s_{n1},s_{n2},...,s_{{\mathrm{nk}}_n}\}\\ k_j=\[\frac{L_j}{\mathrm{\Δ}s}\]\\ j=1,...,n\end{array}\right.$

In formula: S section is gathered；

The kth of j-th strip highway_jIndividual section；

k_jThe section number of j-th strip highway, individual；

L_jThe length of j-th strip highway, km；

Δ s unit space is spaced, km；

N highway bar number, bar.

In the present embodiment, every highway only carries out a flow collection job, therefore every highway is 1 Section.The length of each bar highway is unit space interval, then S={s₁₁s₂₁}。

2.3) prediction period in each section is determined: user is arrived the time of each road segment end and calculated by equation below:

$\left\{\begin{array}{c}{t}_{s_{{\mathrm{jk}}_j}}=t_{s_{j(k_j-1)}}+\frac{\mathrm{\Δ}s}{v_j}\\ t_{s_{j1}}=\frac{t_a+t_b}{2}\end{array}\right.$

In formula:User reaches the kth of j-th strip highway_jThe time of individual road segment end；

User reaches the time of the entrance of j-th strip highway；

v_jThe average speed of operation of j-th strip highway；

t_aThe starting point of the period of entering the station of user's reservation；

t_bThe terminal of the period of entering the station of user's reservation.

In this manner it is possible to judgement sectionPrediction period: ifThen sectionPrediction period be t_i。

In the present embodiment, the time of entering the station is(representing 8: 30, lower same).It is each that user reaches highway 1 The terminal time in section is 9.5, therefore the prediction period of highway 1 is t₉(8:00-9:00) and t₁₀(9:00-10:00).User The terminal time reaching each section of highway 2 is 10.00.Therefore the prediction period of highway 2 is t₉(8:00-9:00) and t₁₀ (9:00-10:00)。

3) volume of traffic in section each to prediction period freeway network is predicted, and specifically includes following steps:

3.1) predict the total wheel traffic in each section according to historical data, use BP neutral net to be predicted, can directly exist Matlab calls the function carried realize.

In the present embodiment, it was predicted that result is as shown in table 5.

Table 5 freeway total wheel traffic predictive value

3.2) reservation traffic total amount is calculatedWith non-reservation traffic total amountIts computing formula is:

$\left\{\begin{array}{c}{Q}_{t_i}=\underset{j}{\Σ}\underset{k_j}{\Σ}{q}_{{\mathrm{jk}}_j}^{t_i}\\ Q_{t_i,reservation}=Q_{t_i}\·\mathrm{\α}\\ Q_{t_i,non-reseration}=Q_{t_i}-Q_{t_i,reservation}\end{array}\right.$

In formula:Prediction period t_iTime road network total wheel traffic；

Prediction period t_iTime road network reservation traffic total amount；

Prediction period t_iTime road network non-reservation traffic total amount；

Prediction period t_iTime j-th strip highway kth_jThe volume of traffic in individual section；

α user app or the market share of wechat public number.

Remaining each parameter meaning is as previously mentioned.

Result of calculation is as shown in table 6.

Table 6 preengages the volume of traffic and non-reservation volume of traffic result of calculation

3.3) according to the historical rethinking rule of vehicle flowrate, non-reservation traffic total amount is allocated in the range of road network, To the non-reservation amount in each section, its computing formula is:

$q_{{\mathrm{jk}}_j,non-reservation}^{t_i}=\frac{q_{{\mathrm{jk}}_j}^{t_i}}{Q_{t_i}}\·Q_{t_i,non-reservation}$

In formula:Prediction period t_iTime j-th strip highway kth_jThe non-reservation traffic in individual section Amount.

Result of calculation is as shown in table 7.

Table 7 non-reservation amount allocation result

3.4) use user equilibrium model that reservation traffic total amount is allocated in the range of road network, obtain the pre-of each section About measure.

$\min Z\left(X\right)=\underset{j}{\mathrm{\Σ}}{\∫}_0^{q_{{\mathrm{jk}}_j,non-reservation}^{t_i}}{t}_{q_{{\mathrm{jk}}_j,non-reservation}^{t_i}}\left(w\right)dw$

$S.t.\left\{\begin{array}{c}\underset{j}{\Σ}{f}_{t_{i}j}=q_{{\mathrm{jk}}_j,non-reservation}^{t_i},\\ f_{t_{i}j}\≥0,\end{array}\right.$

$q_{{\mathrm{jk}}_j,non-reservation}^{t_i}=\underset{j}{\Σ}{f}_{t_{i}j},\∀j$

In formula:Prediction period t_iTime j-th strip highway flow.

Subscriber Impedance Function in road network, its concrete form is:

$t_{q_{{\mathrm{jk}}_j,non-reservation}^{t_i}}\left(w\right)=t_o\[1+\mathrm{\α}{\left(\frac{q_{{\mathrm{jk}}_j,non-reservation}^{t_i}+q_{{\mathrm{jk}}_j,reservation}^{t_i}}{C_{{\mathrm{jk}}_j}}\right)}^{\mathrm{\β}}\]+\frac{c_o(1-\mathrm{\γ})}{\mathrm{\τ}}$

In formula: t_oRunning time under the conditions of freely flowing, min；

The kth of j-th strip highway_jThe road design traffic capacity in individual section, pcu/h；

c_oBasis pass cost use, unit；

The γ margin of preference, %；

The parameter that α, β are to be calibrated；

τ is worth hourage, unit/h.

In the present embodiment, parameter value is α=0.5, β=1.Think that the user walking highway 2 is staggered shifts, at a high speed The non-staggered shifts of user of highway 1.Therefore the impedance function of two highways is respectively as follows:

$t_{q_{11,non-reservation}^{t_i}}\left(w\right)=1.1\×\[1+0.5\left(\frac{q_{{\mathrm{jk}}_j,non-reservation}^{t_i}+q_{{\mathrm{jk}}_j,reservation}^{t_i}}{2100}\right)\]+0.95,$

$t_{q_{21,non-reservation}^{t_i}}\left(w\right)=1.5\×\[1+0.5\left(\frac{q_{{\mathrm{jk}}_j,non-reservation}^{t_i}+q_{{\mathrm{jk}}_j,reservation}^{t_i}}{2000}\right)\]+\mathrm{1.03.}$

Solve linear equation in two unknowns groupObtain reservation amount Allocation result is as shown in table 8.

Table 8 reservation amount allocation result

3.5) using actual reservation amount to be modified the reservation amount in each section, modification rule is as follows:

If 1.ThenConstant.

If 2.Then

In formula:Prediction period t_iTime j-th strip highway kth_jThe actual reservation in individual section is handed over Logical total amount；

Revised result is as shown in table 9:

The correction result of table 9 reservation amount distribution

3.6) each section reservation amount and non-reservation amount are overlapped, obtain the volume of traffic in each section.

$q_{{\mathrm{jk}}_j}^{t_i}=q_{{\mathrm{jk}}_j,reservation}^{t_i}+q_{{\mathrm{jk}}_j,non-reservation}^{t_i}$

In formula:Prediction period t_iTime j-th strip highway kth_jThe reservation volume of traffic in individual section.

Final predicts the outcome as shown in table 10:

Table 10 link counting finally predicts the outcome

4) crowding in each section is predicted: employing saturation is as crowded evaluation index, by each section of threshold decision Congestion state.Specifically include following steps:

4.1) calculating the saturation in each section, computing formula is:

$x_{{\mathrm{jk}}_j}^{t_i}=\frac{q_{{\mathrm{jk}}_j}^{t_i}}{C_{{\mathrm{jk}}_j}}$

In formula:User selects the period t that goes on a journey_iTime j-th strip highway kth_jThe saturation in individual section.

Saturation computation result is as shown in table 11.

Table 11 saturation computation result

4.2) division of crowding grade, is divided into different grades by saturation threshold value by crowding.Can basis The traffic survey data in each city determines threshold value.In the case of without survey data, it is proposed that use table 12 carries out classification.

Table 22

5) programming language is utilized to realize the crowded visualization predicted the outcome output.By different crowding grades with not Same color is corresponding, gives each section when writing code by corresponding color code.Operation code, the most exportable crowded prediction Result.Output result is as shown in Figure 3.

In sum, after using above scheme, the present invention is that the crowded prediction of highway provides new method, it is possible to Adapt to the current new change brought of highway reservation, the most effectively predict the congested conditions of highway, for The induced travel of highway and management work provide more reliable basis, effectively promote the development of China's highway, have Actual promotional value, is worthy to be popularized.

Embodiment described above is only the preferred embodiments of the invention, not limits the practical range of the present invention with this, therefore The change that all shapes according to the present invention, principle are made, all should contain within the scope of the present invention.

Claims (4)

1. the crowded Forecasting Methodology of highway merging historical data and reservation data, it is characterised in that include following step Rapid:

1) highway of selected research, obtains basic data, altogether includes the data of four aspects: highway user reservation number According to, freeway network basic data, freeway traffic flow historical data, expressway tol lcollection policy data；

2) according to highway user reservation data and freeway network basic data, freeway network is carried out space-time and draws Divisional processing also determines the prediction period in each section；

3) volume of traffic in section each to freeway network in prediction period is predicted

First, utilize historical data that traffic total amount is predicted；Then, by total wheel traffic according to reservation amount and non-reservation measure into Non-reservation amount is also modified by row distribution by reservation data；Finally reservation amount after distribution and non-reservation amount are added and obtain each road The volume of traffic of section；

4) crowding in each section is predicted: employing saturation is as crowded evaluation index, by gathering around of each section of threshold decision The state of squeezing；

5) programming language is utilized to realize the crowded visualization predicted the outcome output.

A kind of crowded Forecasting Methodology of highway merging historical data and reservation data the most according to claim 1, its It is characterised by: in step 1) in, described highway user reservation data includes departure place, destination that user inputs, enters the station Period, vehicle, user app or the market share of wechat public number, obtained by user app backstage or wechat public number backstage Take；Described freeway network basic data includes the length of each road, number of track-lines, lane width, design capacity, freedom Time of vehicle operation under the conditions of stream, by Operation and Management of Expressway, department obtains；Described freeway traffic flow history number According to including flow and vehicle speed data, by Operation and Management of Expressway, department obtains；Described expressway tol lcollection policy data bag Including and are worth basic rate, the margin of preference and hourage, by Operation and Management of Expressway, department obtains.

A kind of crowded Forecasting Methodology of highway merging historical data and reservation data the most according to claim 1, its It is characterised by: in step 2) in, according to highway user reservation data and freeway network basic data, to highway Road network carries out space-time division process and determines the prediction period in each section, specifically includes following steps:

2.1) division of time: use between the unit interval consistent with the Time segments division that enters the station on user app or wechat public number Every Δ t, according to 10min, 20min, 30min, the 60min to be asked for of Operation and Management of Expressway department, one day 24 little time-division Become m timeslice；

$\left\{\begin{array}{c}T=\{t_1,...,t_i,...,t_m\}\\ m=\frac{1440}{\mathrm{\Δ}t}\\ i=1,...,m\end{array}\right.$

In formula: T timeslice set；

t_iT_iIndividual timeslice；

M timeslice number；

Δ t unit interval, unit of time is min；

2.2) division in space: use the pavement section gathering traffic flow data with Operation and Management of Expressway department consistent Unit space interval delta s, according to the 1km to be asked for of Operation and Management of Expressway department, 2km, 5km, 10km etc., by road network All highways are divided intoIndividual section；

$\left\{\begin{array}{c}S=\{s_{11},s_{12},...,s_{1k_1},...,s_{j1},s_{j2},...,s_{{\mathrm{jk}}_j},...,s_{n1},s_{n2},...,s_{{\mathrm{nk}}_n}\}\\ k_j=\[\frac{L_j}{\mathrm{\Δ}s}\]\\ j=1,...,n\end{array}\right.$

In formula: S section is gathered；

The kth of j-th strip highway_jIndividual section；

k_jThe section number of j-th strip highway；

L_jThe length of j-th strip highway, unit is km；

Δ s unit space is spaced, and unit is km；

N highway bar number；

2.3) prediction period in each section is determined: user is arrived the time of each road segment end and calculated by equation below:

$\left\{\begin{array}{c}{t}_{s_{{\mathrm{jk}}_j}}=t_{s_{j(k_j-1)}}+\frac{\mathrm{\Δ}s}{v_j}\\ t_{s_{j1}}=\frac{t_a+t_b}{2}\end{array}\right.$

In formula:User reaches the kth of j-th strip highway_jThe time of individual road segment end；

User reaches the time of the entrance of j-th strip highway；

v_jThe average speed of operation of j-th strip highway；

t_aThe starting point of the period of entering the station of user's reservation；

t_bThe terminal of the period of entering the station of user's reservation；

So, it becomes possible to judge sectionPrediction period: ifThen sectionPrediction period be t_i；

In step 3) in, the volume of traffic in section each to prediction period freeway network is predicted, and specifically includes following steps:

3.1) predict the total wheel traffic in each section according to historical data, use BP neutral net to be predicted, it is possible to directly to exist Matlab calls the function carried realize；

3.2) reservation traffic total amount is calculatedWith non-reservation traffic total amountIts computing formula is:

$\left\{\begin{array}{c}{Q}_{t_i}=\underset{j}{\Σ}\underset{k_j}{\Σ}{q}_{{\mathrm{jk}}_j}^{t_i}\\ Q_{t_i,reservation}=Q_{t_i}\·\mathrm{\α}\\ Q_{t_i,non-reseration}=Q_{t_i}-Q_{t_i,reservation}\end{array}\right.$

In formula:Prediction period t_iTime road network total wheel traffic；

Prediction period t_iTime road network reservation traffic total amount；

Prediction period t_iTime road network non-reservation traffic total amount；

Prediction period t_iTime j-th strip highway kth_jThe volume of traffic in individual section；

α user app or the market share of wechat public number；

Remaining each parameter meaning is as previously mentioned；

3.3) according to the historical rethinking rule of vehicle flowrate, non-reservation traffic total amount is allocated in the range of road network, obtains each The non-reservation amount in section, its computing formula is:

$q_{{\mathrm{jk}}_j,non-reservation}^{t_i}=\frac{q_{{\mathrm{jk}}_j}^{t_i}}{Q_{t_i}}\·Q_{t_i,non-reservation}$

In formula:Prediction period t_iTime j-th strip highway kth_jThe non-reservation volume of traffic in individual section；

3.4) use user equilibrium model that reservation traffic total amount is allocated in the range of road network, obtain the reservation in each section Amount；

$\min Z\left(X\right)=\underset{j}{\Σ}{\∫}_0^{q_{{\mathrm{jk}}_j,non-reservation}^{t_i}}{t}_{q_{{\mathrm{jk}}_j,non-reservation}^{t_i}}\left(w\right)dw$

$S.t.\left\{\begin{array}{c}\underset{j}{\Σ}{f}_{t_{i}j}=q_{{\mathrm{jk}}_j,non-reservation}^{t_i},\\ f_{t_{i}j}\≥0,\end{array}\right.$

$q_{{\mathrm{jk}}_j,non-reservation}^{t_i}=\underset{j}{\Σ}{f}_{t_{i}j},\∀j$

In formula:Prediction period t_iTime_jThe flow of bar highway；

Subscriber Impedance Function in road network, its concrete form is:

$t_{q_{{\mathrm{jk}}_j,non-resevation}^{t_i}}\left(w\right)=t_o\[1+\mathrm{\α}{\left(\frac{q_{{\mathrm{jk}}_j,non-reservation}^k+q_{{\mathrm{jk}}_j,reservation}^{t_i}}{C_{{\mathrm{jk}}_j}}\right)}^{\mathrm{\β}}\]+\frac{c_o(1-\mathrm{\γ})}{\mathrm{\τ}}$

In formula: t_oRunning time under the conditions of freely flowing, unit is min；

The_jThe kth of bar highway_jThe road design traffic capacity in individual section, unit is pcu/h；

c_oBasis pass cost use, unit is unit；

The γ margin of preference, unit is %；

The parameter that α, β are to be calibrated；

τ is worth hourage, and unit is unit/h；

3.5) using actual reservation amount to be modified the reservation amount in each section, modification rule is as follows:

If 1.ThenConstant；

If 2.Then

In formula:Prediction period t_iTime_jThe kth of bar highway_jThe actual reservation traffic in individual section is total Amount；

3.6) each section reservation amount and non-reservation amount are overlapped, obtain the volume of traffic in each section；

$q_{{\mathrm{jk}}_j}^{t_i}=q_{{\mathrm{jk}}_j,reservation}^{t_i}+q_{{\mathrm{jk}}_j,non-reservation}^{t_i}$

In formula:Prediction period t_iTime_jThe kth of bar highway_jThe reservation volume of traffic in individual section；

In step 4) in, it was predicted that the crowding in each section, specifically include following steps:

4.1) calculating the saturation in each section, computing formula is:

$x_{{\mathrm{jk}}_j}^{t_i}=\frac{q_{{\mathrm{jk}}_j}^{t_i}}{C_{{\mathrm{jk}}_j}}$

In formula:Prediction period t_iTime_jThe kth of bar highway_jThe saturation in individual section；

4.2) division of crowding grade, is divided into different grades by saturation threshold value, it is possible to according to each city by crowding The traffic survey data in city determines threshold value.

A kind of crowded Forecasting Methodology of highway merging historical data and reservation data the most according to claim 1, its It is characterised by: in step 5) in, utilize programming language to realize the crowded visualization predicted the outcome output, by different crowded Degree grade is corresponding from different colors, when writing code by each section of corresponding color code imparting, and operation code, can be defeated Go out crowded predicting the outcome.