CN104376716A - Method for dynamically generating bus timetables on basis of Bayesian network models - Google Patents

Abstract

The invention relates to a method for dynamically generating bus timetables on the basis of Bayesian network models. The method includes screening microscopic and macroscopic factors affecting dynamic generation of the bus timetables; building the double-layer microscopic and macroscopic Bayesian network models for dynamically generating the bus timetables, and in other words, building the Bayesian network models for forecasting dynamic variation of bus environments and the Bayesian network models for dynamically generating the bus timetables; predicting transport capacity and transport volume occurrence probabilities of various routes under the condition of random disturbance and reasons for unbalance of the transport capacity and the transport volumes of the various routes; combining scheduling policies with one another and generating possible timetable schemes around the target for timely evacuating passengers; computing various indexes for evaluating the quality of the timetables from the points of governments, enterprises and the passengers and evaluating the quality of the timetables. The method has the advantages that a function of dynamically adjusting the timetables according to variation of the bus environments can be implemented, and accordingly technical support can be provided for daily bus operation management.

Description

A kind of transit scheduling dynamic creation method based on Bayesian network model

Technical field

The present invention relates to Bus transit informization technical field, specifically a kind of transit scheduling dynamic creation method based on Bayesian network model.

Background technology

Transit scheduling establishment is one of core missions of the daily operation of public transport, and according to resident trip spatial-temporal distribution characteristic, rationalization arranges departure frequency and the type thereof of day part, mainly solves the maximum matching problem of transport power and freight volume.When in reality enchancement factor interference cause bus passenger flow or running time change time, this cause public transport transport power and freight volume unbalance, thus bus dispatching scheme lost efficacy.Therefore, according to public transport environment dynamic change dynamic conditioning timetable, theory value and realistic meaning is had.

The factor that direct decision transit scheduling lost efficacy is transport power and the freight volume of up-downlink direction, they are by external environment influence such as Changes in weather, traffic congestion, large-scale activities, when detecting traffic events, assess it and affect passenger flow or running time intensity of variation respectively, and then analyze the failure cause of timetable, dynamic conditioning timetable accordingly.At present, numerous Chinese scholars is according to public transport environment dynamic change dynamic conditioning timetable, and main Research Thinking has two:

One, passenger flow forecast or running time change, on the one hand, utilize the method such as multiple linear regression, the equation of structure, and qualitative, quantitative further investigated affect relevance between many factors that passenger flow or running time change, line correlation sensitivity analysis of going forward side by side; On the other hand, utilize time series method, be regarded as a black box, directly disclose the evolution trend of passenger flow or running time change, for establishment timetable provides data supporting.

Two, working out timetable, on the one hand, above-mentioned working foundation is studied the statistical law of passenger flow and running time, when detecting traffic events, building timetable compiling model, utilize Optimum Theory to generate timetable; On the other hand, utilize manual simulation's technology such as neural network, the thoughtcast of operation simulation personnel, environmentally changes, adjustment time adjustment table.

As from the foregoing, existing research way cannot solve the transit scheduling dynamic change chain reaction process that random disturbance causes, should from entirety, announcement affects external environment condition change and how to cause passenger flow or running time change, and then affect timetable dynamic adjustment process and how to occur, and mutually cause between them, interfere, transform and the complex relationship such as coupling, predict that the transit scheduling under complicated traffic environment change situation generates and probability of happening.

Bayesian network is that one portrays causal probability graph model between things, and the very applicable generation to accident and the chain reaction process caused thereof carry out modeling analysis.Based on this, unbalance reason between the transport power that analyzing influence transit scheduling of the present invention dynamically generates and freight volume, external environmental factor is considered as input, analyze it and how to affect passenger flow or running time change, and then how to cause between public transport transport power and freight volume unbalance, according to existing conveyance equilibrium, output is the result of the maximum coupling freight volume of timetable, control inputs can the change of control section state, the each external environment condition node built accordingly in accident Bayesian network inputs four etale topology network structures of passenger flow or the output of running time transport power freight volume calculating timetable decision node, realize the transit scheduling under prediction complicated traffic environment change situation and probability of happening thereof, for bus dynamic dispatching provides reliable technical support.

Summary of the invention

The invention provides a kind of transit scheduling dynamic creation method based on Bayesian network model, on the various influence factor bases of analyzing influence transit scheduling dynamic change, in conjunction with actual public transport dynamic data, portray the cause-effect relationship between them, when changing according to intelligent public transportation dispatching detection of platform external environment condition, transport power under the various complicated traffic environment of reasoning and the unbalance reason of freight volume and probability of happening thereof, calculate departure frequency and scheduling type thereof accordingly, the present invention is mainly used in generating timetable according to public transport external environment condition change tread, for the daily operation management of public transport provides technical support.

The present invention program is achieved through the following technical solutions:

The invention provides a kind of transit scheduling dynamic creation method based on Bayesian network model, comprise the following steps:

(1) the numerous external environmental factors quantitatively affecting transit scheduling with the qualitative method screening combined and dynamically generate are adopted, comprising traffic hazard causes running time to change, and then cause transport power deficiency or large-scale activity to cause passenger flow to fluctuate, and then cause freight volume not enough;

(2) the upper and lower two-layer Bayesian network model that transit scheduling dynamically generates is built, wherein: upper layer model portrays the external environment condition random disturbance causing passenger flow or running time change, it is the initial conditions of underlying model, underlying model describes and causes transport power and the unbalance external environment condition random disturbance of freight volume, provides data supporting for timetable generates;

(3) when traffic events occurs, the realtime running data of combined with intelligent bus dispatching platform, the various circuit transport power freight volume probability of happening utilizing step (2) micro-macromodel to predict and unbalance reason, in conjunction with scheduling strategy, evacuate passenger in time for target around with minimum cost, generate multiple timetable scheme;

(4) passenger's Waiting time, the website of often kind of timetable scheme in calculation procedure (3) are detained situation, operation cost indices, assess its quality.

Improve as one, adopt quantitatively and the qualitative method combined, screening affects the many factors process that transit scheduling dynamically generates, and comprising:

(1) from microcosmic, N number of external environment condition enchancement factor X=(X is analyzed ₁, X ₂..., X _n) how to disturb passenger flow fluctuation pf at time t _t(X) or running time change pt _t(X), as road type, road conditions, traffic hazard, large-scale activity, traffic control and Changes in weather etc., so affect needed for timetable with car demand and available vehicle number;

(2) from macroscopically, the up direct determinative of circuit that current T period timetable generates is disclosed with line downstream direct acting factor the up car demand of current T period descending car demand up available vehicle number with descending available vehicle number and within the lower T+1 period up car demand descending car demand up available vehicle number with descending available vehicle number

As further improvement, according to the microcosmic affecting transit scheduling generation filtered out and Macroscopic Factors, build two layers of Bayesian network model that transit scheduling dynamically generates, comprising:

(1) node abstraction definition,

In public transport environment dynamic change forecast Bayesian network model, its condition node is N number of external environment condition enchancement factor X=(X ₁, X ₂..., X _n), comprise road type, road conditions, traffic hazard, large-scale activity, traffic control and Changes in weather etc.; Its decision node is passenger flow fluctuation or the running time change Y={pf of time t _t(X), pt _t(X) }.

Dynamically generate in Bayesian network model at transit scheduling, its condition node is on circuit or down direction uses car demand within current T period and next T+1 period, and their available vehicle number, namely $Z=\{{\mathrm{pr}}_T^U\left(X\right),{\mathrm{pr}}_T^D\left(X\right),{\mathrm{pv}}_T^U\left(X\right),{\mathrm{pv}}_T^D\left(X\right);{\mathrm{pr}}_{T+1}^U\left(X\right),{\mathrm{pr}}_{T+1}^D\left(X\right),{\mathrm{Pv}}_{T+1}^U\left(X\right),{\mathrm{pv}}_{T+1}^D\left(X\right)\};$ Its decision node is departure frequency on current T period circuit or descending

(2) Structure learning,

Utilize the conditional independence method of inspection, respectively the dynamic change forecast of public transport environment and transit scheduling are dynamically generated to all nodes of Bayesian network model, if any two nodes and between interdepend, there is directed edge to be connected, build a directed acyclic graph, set up their bayesian network structure figure S.

(3) parameter learning,

Utilize maximum Likelihood, respectively dynamically Bayesian network model is generated to the dynamic change forecast of public transport environment and transit scheduling, at they given network topology structure S and training sample set D separately, utilize priori, determine that the conditional probability density at each node place of respective Bayesian network model is:

Describe external environment condition change and passenger flow or running time fluctuation between probability causal relation

$p(Y/X)=\frac{p\left(Y\right)\×p\left(X\right|Y)}{p\left(X\right)};$

State transfer relationship between the departure frequency of portraying external environment condition random disturbance and circuit up-downgoing

$p(S/Z)=\frac{p\left(S\right)\×p\left(Z\right|S)}{p\left(Z\right)}.$

Preferred as one, the various circuit transport power freight volume probability of happening under prediction random disturbance and their unbalance reasons, comprising:

(1) for certain circuit, total K website and M car, utilize the real data of intelligent public transportation dispatching platform, in conjunction with the latitude and longitude coordinates of every i car, estimates that the passenger flow arriving k website at time t arrives number and this vehicle arrives the running time at first and last station

(2) when traffic events being detected, N number of external environment condition enchancement factor X=(X is determined ₁, X ₂..., X _n) value, utilize clique tree propagation algorithm, according to public transport environment dynamic change model X → Y={pf _t(X), pt _t(X) their wave time pt }, is predicted _t(X) and change passenger flow pf _tand their probability of happening (X);

(3) on the basis of the above, gather circuit up-downgoing and use car demand in certain T period ${\mathrm{pr}}_T^U\left(X\right)=\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\left[p_{t-{\mathrm{at}}_t^i-{\mathrm{pt}}_t\left(X\right)}^k{+\mathrm{pf}}_{t-{\mathrm{at}}_t^i-{\mathrm{pt}}_t\left(X\right)}\left(X\right)\right]{,\mathrm{pr}}_T^D\left(X\right)=\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\underset{i=1}{\overset{M}{\mathrm{\Σ}}}[p_{t-{\mathrm{at}}_t^i-{\mathrm{pt}}_t\left(X\right)}^k+{\mathrm{pf}}_{t-{\mathrm{at}}_t^i-{\mathrm{pt}}_t\left(X\right)}\left(X\right)],$ Available vehicle number ${\mathrm{pv}}_T^U\left(X\right)=\underset{i=1}{\overset{M}{\mathrm{\Σ}}}f({\mathrm{at}}_t^i+{\mathrm{pt}}_t\left(X\right)),{\mathrm{pv}}_T^D\left(X\right)=\underset{i=1}{\overset{M}{\mathrm{\Σ}}}f({\mathrm{at}}_t^i+{\mathrm{pt}}_t\left(X\right))$ And their probability, analyze between transport power and freight volume and whether mate, for transit scheduling establishment provides data supporting.

Preferential as another kind, in conjunction with scheduling strategy, around evacuating passenger's target in time, generating possible timetable scheme, comprising:

(1) adopt even equilibrium to dispatch a car strategy, consider single scheduling method, meeting capacity consistency c, the degree of crowding γ ∈ [γ of bus _min, γ _max], on government departure interval F basis, determine departure frequency scope with ${\mathrm{Trip}}_T^D\left(X\right)=\max (\frac{{\mathrm{pr}}_T^D\left(X\right)}{\mathrm{\γ}\·c},F)$ And their probability;

(2) according to unbalance between transport power freight volume, utilize clique tree propagation algorithm, utilize timetable to generate Bayesian network model $Z=\{{\mathrm{pr}}_T^U\left(X\right),{\mathrm{pr}}_T^D\left(X\right),{\mathrm{pv}}_T^U\left(X\right),{\mathrm{pv}}_T^D\left(X\right);{\mathrm{pr}}_{T+1}^U\left(X\right),{\mathrm{pr}}_{T+1}^D\left(X\right),{\mathrm{pv}}_{T+1}^U\left(X\right),{\mathrm{pv}}_{T+1}^D\left(X\right)\}\→$ $S=\{{\mathrm{Trip}}_T^D\left(X\right),{\mathrm{Trip}}_T^U\left(X\right)\},$ Generate possible timetable scheme.

As preferential further, from government, enterprise and passenger's angle, calculate the various indexs of assessment timetable quality, assess their quality, comprising:

(1) on timetable scheme basis, consider unit mileage running cost l, calculate total passenger's Waiting time of often kind of scheme $\underset{k=1}{\overset{K}{\mathrm{\Σ}}}[\frac{T\·\underset{T}{\∫}{p}_t^k\mathrm{dt}}{{\mathrm{Trip}}_T^U\left(X\right)}+\frac{T\·\underset{T}{\∫}{p}_t^k\mathrm{dt}}{{\mathrm{Trip}}_T^D\left(X\right)}],$ Website is detained situation $\underset{k=1}{\overset{K}{\mathrm{\Σ}}}[{\mathrm{Trip}}_T^U\left(X\right)\·\mathrm{\γ}\·c-\underset{T}{\∫}{p}_t^k\mathrm{dt}],$ Operation cost $\left[\right[{\mathrm{Trip}}_T^U\left(X\right)+{\mathrm{Trip}}_T^D\left(X\right)]\·l$ Deng indices;

(2) from government, enterprise and passenger's angle, the quality of the above-mentioned each timetable scheme of comprehensive evaluation, selects the best time to show scheme for public traffic management department and provides decision support.

The present invention improves owing to have employed above-mentioned several measure, utilize Bayesian network to portray external environment condition change and how to cause passenger flow or running time change, and then affect the process of dynamic conditioning of transit scheduling, avoid existing method and cannot solve the transport power and the unbalance chain reaction process of freight volume that accident causes, the change of public transport external environment condition can be excavated from raw sample data, passenger flow or running time fluctuation, coupled relation between timetable dynamic conditioning, from in advance, thing neutralizes the reason and development trend thereof how the multi-faceted real-time analysis traffic events of overall process afterwards to cause transit scheduling to change, for bus dynamic dispatching provides data supporting.

Accompanying drawing explanation

Fig. 1 is the structural representation that the transit scheduling that the present invention relates to generates Bayesian network;

Fig. 2 is process flow diagram of the invention process.

Embodiment

Be described further below in conjunction with accompanying drawing provided by the present invention:

As shown in Figure 1, the invention provides a kind of transit scheduling dynamic creation method based on Bayesian network model, the process occurring according to traffic events, develop and develop, two layers of microcosmic that structure transit scheduling dynamically generates and macroscopical Bayesian network model, unbalance reason between the transport power that analyzing influence transit scheduling dynamically generates and freight volume, external environmental factor is considered as input, analyze it and how to affect passenger flow or running time change, and then how causing between public transport transport power and freight volume unbalance, output is the result of the maximum coupling freight volume of timetable.Control inputs can the change of control section state, the each external environment condition node built accordingly in accident Bayesian network inputs four etale topology network structures of passenger flow or the output of running time prediction transport power freight volume calculating timetable decision node, realize the transit scheduling under prediction complicated traffic environment change situation and probability of happening thereof, for bus dynamic dispatching provides reliable technical support.

As shown in Figure 2, the invention provides a kind of transit scheduling dynamic creation method based on Bayesian network model, comprise four steps such as Analysis on Mechanism, modelling, modelling verification and model analysis utilization, embodiment is as follows.

Step 1: Analysis on Mechanism, adopts quantitatively and the qualitative method that combines screens the numerous external environmental factors affecting transit scheduling and dynamically generate, and sets up the factor storehouse affecting transit scheduling and lost efficacy.

Step 1.1: from microcosmic, analyzes N number of external environment condition enchancement factor X=(X ₁, X ₂..., X _n) how to disturb passenger flow fluctuation pf at time t _t(X) or running time change pt _t(X), as road type, road conditions, traffic hazard, large-scale activity, traffic control and Changes in weather etc., so affect needed for timetable with car demand and available vehicle number;

Step 1.1.1: invite expert to have an informal discussion, select all n possibility influence factors of passenger flow fluctuation or running time change respectively, the former relates to season, festivals or holidays, period, large-scale activity, traffic control, vehicle trouble, weather etc., and the latter contains road type, traffic flow, traffic congestion, type of site, passenger flow and weather etc.

Step 1.1.2: combined with intelligent bus dispatching platform, the numerical value a of Dynamic Acquisition j external environment influence factor i at any time _ij, and their corresponding passenger flows or running time y _i, will altogether m bar data record as sample D, formation condition matrix A=(a _ij) _mnwith decision vector Y=(y _i) _m.

Step 1.1.3: according to AB=Y, based on least square method, calculates B=(b ₁, b ₂..., b _n)=(A ' A) ^-1(A ' Y), right if b _jbe greater than artificial threshold values σ, this factor determines passenger flow or running time, obtains n influence factor variable.

Step 1.2: from macroscopically, disclose the up direct determinative of circuit that current T period timetable generates with line downstream direct acting factor the up car demand of current T period descending car demand up available vehicle number with descending available vehicle number and within the lower T+1 period up car demand descending car demand up available vehicle number with descending available vehicle number

Step 1.2.1: according to the service time scope of circuit, be divided into N number of period, calculated up-downlink direction respectively and use car demand within current T period and next T+1 period and their available vehicle number ${\mathrm{pv}}_T^U\left(X\right),{\mathrm{pv}}_T^D\left(X\right).$

Step 1.2.2: combined with intelligent bus dispatching platform, for certain circuit, total K website and M car, in conjunction with the latitude and longitude coordinates of every i car, estimate that the passenger flow arriving k website at time t arrives number and this vehicle arrives the running time at first and last station

Step 1.2.3: gather circuit and use car demand in the T period ${\mathrm{pr}}_T^U\left(X\right)=\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\left[p_{t-{\mathrm{at}}_t^i-{\mathrm{pt}}_t\left(X\right)}^k{+\mathrm{pf}}_{t-{\mathrm{at}}_t^i-{\mathrm{pt}}_t\left(X\right)}\left(X\right)\right],$ ${\mathrm{pr}}_T^D\left(X\right)=\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\underset{i=1}{\overset{M}{\mathrm{\Σ}}}[p_{t-{\mathrm{at}}_t^i-{\mathrm{pt}}_t\left(X\right)}^k+{\mathrm{pf}}_{t-{\mathrm{at}}_t^i-{\mathrm{pt}}_t\left(X\right)}\left(X\right)],$ Available vehicle number ${\mathrm{pv}}_T^U\left(X\right)=\underset{i=1}{\overset{M}{\mathrm{\Σ}}}f\left({\mathrm{at}}_t^i{+\mathrm{pt}}_t\left(X\right)\right),$ analyze between transport power and freight volume and whether mate, for transit scheduling establishment provides data supporting.

Step 2: modelling, build two layers of microcosmic and macroscopical Bayesian network model up and down that transit scheduling dynamically generates, comprise node variable definition, determine the span of each condition and decision variable and prior probability distribution, Structure learning and parameter learning four part, its at the middle and upper levels model portray cause passenger flow or running time change external environment condition random disturbance, it is the initial conditions of underlying model, underlying model describes and causes transport power and the unbalance external environment condition random disturbance of freight volume, provides data supporting for timetable generates; .

Step 2.1: variable node defines,

In public transport environment dynamic change forecast Bayesian network model, total n+2 node variable X={X ₁, X ₂..., X _n∪ Y={pf _t(X), pt _t(X) n condition and 2 decision variable nodes }, are divided into.The former is public transport external environment condition random disturbance input key element, and the latter is passenger flow or running time Output rusults, pay close attention between public transport external environment condition random disturbance input key element and influence each other, and how their change causes passenger flow or running time change.

Dynamically generate in Bayesian network model at transit scheduling, its condition node is on circuit or down direction uses car demand within current T period and next T+1 period, and their available vehicle number, namely $Z=\{{\mathrm{pr}}_T^U\left(X\right),{\mathrm{pr}}_T^D\left(X\right),{\mathrm{pv}}_T^U\left(X\right),{\mathrm{pv}}_T^D\left(X\right);{\mathrm{pr}}_{T+1}^U\left(X\right),{\mathrm{pr}}_{T+1}^D\left(X\right),{\mathrm{pv}}_{T+1}^U\left(X\right),{\mathrm{pv}}_{T+1}^D\left(X\right)\};\→$ Its decision node is departure frequency on current T period circuit or descending

Step 2.2: determine the above-mentioned condition of microcosmic and macromodel and the span of decision variable node and prior probability distribution between them respectively.

Step 2.2.1: any node micro-(grand) being seen to model the span of (S ∪ Z) by its discretize K feature value state space be $x_i^j=X_{\min }^i+(j-1)*[X_{\max }^i-X_{\min }^i]/K.$

Step 2.2.2: add up any node that micro-(grand) sees model (S ∪ Z) value state probability $p\left(x_i^j\right)=\mathrm{Count}\left(x_i^j\right)/\underset{j=1}{\overset{K}{\mathrm{\Σ}}}\mathrm{Count}\left(x_i^j\right),$ And Two Variables (S ∪ Z) and (S ∪ Z) prior probability distribution between their state values different the wherein number of times that occurs at sample set D of Count () presentation of events.

Step 2.3: build topological structure between microcosmic and each node of macroscopical Bayesian network respectively.

Step 2.3.1: adopt K2 algorithm, training set carries out unsupervised machine learning, obtains the initial network structure that micro-(grand) sees model respectively.

Step 2.3.2: the priori utilizing expert, based on the conditional independence method of inspection, if micro-(grand) sees any two nodes of model (S ∪ Z) and interdepend between (S ∪ Z), there is directed edge and be connected the network structure of micro-(grand) model is finely tuned.

Step 2.3.3: detect micro-(grand) sight prototype network structure after obtaining adjustment and whether meet the requirements, if meet the demands, export the bayesian network structure figure S of public transport dynamic environment forecast (timetable generation); Otherwise return step 2.3.2, continue the network structure that micro-(grand) sees model.

Step 2.4: on above-mentioned network structure basis, utilize maximum Likelihood, to estimate in microcosmic and macromodel conditional probability distribution table between each node.

Step 2.4.1: to any node of micro-macromodel (S ∪ Z), combines prior distribution and likelihood function, estimated parameter

Step 2.4.2: suppose that θ is the stochastic distribution of Dirichlet function, makes the likelihood function of θ be $=\underset{k=1}{\overset{r_i}{\mathrm{\Π}}}\underset{i=1}{\overset{n}{\mathrm{\Π}}}P\left(X_i\right|{\mathrm{\θ}}_i)=\underset{i=1}{\overset{n}{\mathrm{\Π}}}\underset{j=1}{\overset{K}{\mathrm{\Π}}}P\left(x_i^j\right|{\mathrm{\θ}}_{\mathrm{ij}}),$ Because $L\left({\mathrm{\θ}}_{\mathrm{ij}}\right|X)=P\left(X\right|{\mathrm{\θ}}_{\mathrm{ij}})=\underset{k=1}{\overset{r_i}{\mathrm{\Π}}}{{\mathrm{\θ}}_{\mathrm{ijk}}}^{N_{\mathrm{ijk}}},$ According to $\frac{\∂\left(L\left(\mathrm{\θ}\right|X)\right)}{\∂\mathrm{\θ}}=0,$ Derive thus estimate any node of micro-macromodel p (X between (S ∪ Z) _i| X _n+1).Wherein: r _ifor X _ieigenwert number; N _ijkfor nodes X _iwhen getting kth eigenwert, in the quantity of a father node value jth eigenwert.

Step 2.4.3: according to above-mentioned formulation process, conditional probability distribution table between each node can be calculated in microcosmic and macromodel, conditional probability between i.e. public transport external environment condition random disturbance key element and passenger flow or running time fluctuation, and then affect transport power and freight volume and to fluctuate conditional probability between (timetable dynamic conditioning).

Describe external environment condition change and passenger flow or running time fluctuation between probability causal relation:

$p(Y/X)=\frac{p\left(Y\right)\×p\left(X\right|Y)}{p\left(X\right)};$

State transfer relationship between the departure frequency of portraying external environment condition random disturbance and circuit up-downgoing:

$p(S/Z)=\frac{p\left(S\right)\×p\left(Z\right|S)}{p\left(Z\right)}.$

Step 3: using part training set as test sample book, the precision of microcosmic and macromodel in inspection, if model result does not reach re-set target, returns step 2.

Step 4: model analysis is used, the transit scheduling dynamic generative process of the various complicated traffic environment of Inference Forecast.

Step 4.1: combined with intelligent bus dispatching platform, when the change of public transport external environment condition, the measured value x=(x of each influence factor of dynamic monitoring ₁, x ₂..., x _n), right calculate its eigenstate if (x _i? with between), determine current all node states of network model

Step 4.2: utilize clique tree propagation algorithm, according to microvisual model X → Y={pf _t(X), pt _t(X) }, for K website and M the car of certain circuit, in conjunction with the latitude and longitude coordinates of every i car, estimate that the passenger flow arriving k website at time t arrives number and this vehicle arrives the running time at first and last station

Step 4.3: on the basis of the above, gathers circuit up-downgoing and uses car demand in certain T period ${\mathrm{pr}}_T^U\left(X\right)=\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\underset{i=1}{\overset{M}{\mathrm{\Σ}}}\left[p_{t-{\mathrm{at}}_t^i-{\mathrm{pt}}_t\left(X\right)}^k{+\mathrm{pf}}_{t-{\mathrm{at}}_t^i-{\mathrm{pt}}_t\left(X\right)}\left(X\right)\right]{,\mathrm{pr}}_T^D\left(X\right)=\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\underset{i=1}{\overset{M}{\mathrm{\Σ}}}[p_{t-{\mathrm{at}}_t^i-{\mathrm{pt}}_t\left(X\right)}^k+{\mathrm{pf}}_{t-{\mathrm{at}}_t^i-{\mathrm{pt}}_t\left(X\right)}\left(X\right)],$ Available vehicle number ${\mathrm{pv}}_T^U\left(X\right)=\underset{i=1}{\overset{M}{\mathrm{\Σ}}}f({\mathrm{at}}_t^i+{\mathrm{pt}}_t\left(X\right)),{\mathrm{pv}}_T^D\left(X\right)=\underset{i=1}{\overset{M}{\mathrm{\Σ}}}f({\mathrm{at}}_t^i+{\mathrm{pt}}_t\left(X\right))$ And their probability of happening, analyzing between transport power and freight volume and whether mate, providing data supporting for analyzing the unbalance reason of transit scheduling.

Step 4.4: adopt even equilibrium to dispatch a car strategy, consider single scheduling method, meeting capacity consistency c, the degree of crowding γ ∈ [γ of bus _min, γ _max], on government departure interval F basis, tentatively determine departure frequency scope with and its probability, utilize clique tree propagation algorithm, according to $Z=\{{\mathrm{pr}}_T^U\left(X\right),{\mathrm{pr}}_T^D\left(X\right),{\mathrm{pv}}_T^U\left(X\right),{\mathrm{pv}}_T^D\left(X\right);{\mathrm{pr}}_{T+1}^U\left(X\right),{\mathrm{pr}}_{T+1}^D\left(X\right),{\mathrm{pv}}_{T+1}^U\left(X\right),{\mathrm{pv}}_{T+1}^D\left(X\right)\}\→$ $S=\{{\mathrm{Trip}}_T^D\left(X\right),{\mathrm{Trip}}_T^U\left(X\right)\},$ Generate possible timetable scheme.

Step 4.5: on timetable scheme basis, considers unit mileage running cost l, calculates total passenger's Waiting time of often kind of scheme $\underset{k=1}{\overset{K}{\mathrm{\Σ}}}[\frac{T\·\underset{T}{\∫}{p}_t^k\mathrm{dt}}{{\mathrm{Trip}}_T^U\left(X\right)}+\frac{T\·\underset{T}{\∫}{p}_t^k\mathrm{dt}}{{\mathrm{Trip}}_T^D\left(X\right)}],$ Website is detained situation $\underset{k=1}{\overset{K}{\mathrm{\Σ}}}[{\mathrm{Trip}}_T^U\left(X\right)\·\mathrm{\γ}\·c-\underset{T}{\∫}{p}_t^k\mathrm{dt}],$ Operation cost etc. index, from government, enterprise and passenger's angle, the quality of the above-mentioned each timetable scheme of comprehensive evaluation, and carry out its timetable of backward reasoning and to lose efficacy possible reason of changes, select the best time to show scheme for public traffic management department and decision support is provided.

What more than enumerate is only specific embodiments of the invention.Obviously, the invention is not restricted to above embodiment, many distortion can also be had, as: the present invention changes structural design and the parametric learning method of Bayesian network, the influence factor affecting passenger flow or running time can be expanded, as: road types etc., use different scheduling strategies and appraisal procedure.All distortion that those of ordinary skill in the art can directly derive from content disclosed by the invention or associate, all should think protection scope of the present invention.

Claims (6)

1. the present invention relates to a kind of transit scheduling dynamic creation method based on Bayesian network model, it is characterized in that, comprise the following steps:

(1) the numerous external environmental factors quantitatively affecting transit scheduling with the qualitative method screening combined and dynamically generate are adopted, comprising traffic hazard causes running time to change, and then cause transport power deficiency or large-scale activity to cause passenger flow to fluctuate, and then cause freight volume not enough;

(2) the upper and lower two-layer Bayesian network model that transit scheduling dynamically generates is built, wherein: upper layer model portrays the external environment condition random disturbance causing passenger flow or running time change, it is the initial conditions of underlying model, underlying model describes and causes transport power and the unbalance external environment condition random disturbance of freight volume, provides data supporting for timetable generates;

(3) when traffic events occurs, the realtime running data of combined with intelligent bus dispatching platform, the various circuit transport power freight volume probability of happening utilizing step (2) micro-macromodel to predict and unbalance reason, in conjunction with scheduling strategy, evacuate passenger in time for target around with minimum cost, generate multiple timetable scheme;

(4) passenger's Waiting time, the website of often kind of timetable scheme in calculation procedure (3) are detained situation, operation cost indices, assess its quality.

2. a kind of transit scheduling dynamic creation method based on Bayesian network model according to claim 1, it is characterized in that: the quantitative and qualitative method combined in described step (1), comprise from microcosmic, analyze N number of external environment condition enchancement factor X=(X ₁, X ₂..., X _n) how to disturb passenger flow fluctuation pf at time t _t(X) or running time change pt _t(X); From macroscopically, disclose the up direct determinative of circuit that current T period timetable generates with line downstream direct acting factor the up car demand of current T period descending car demand up available vehicle number with descending available vehicle number and within the lower T+1 period up car demand descending car demand up available vehicle number with descending available vehicle number

3. the transit scheduling dynamic creation method based on Bayesian network model according to claim 2, it is characterized in that: what step (1) filtered out by described step (2) affect microcosmic that transit scheduling generates and Macroscopic Factors, build public transport environment dynamic change Bayesian network model X → Y={pf respectively _t(X), pt _t(X) } and transit scheduling generate Bayesian network model $Z=\{{\mathrm{pr}}_T^U\left(X\right),{\mathrm{pr}}_T^D\left(X\right),{\mathrm{pv}}_T^U\left(X\right),{\mathrm{pv}}_T^D\left(X\right);{\mathrm{pr}}_{T+1}^U\left(X\right),{\mathrm{pr}}_{T+1}^D\left(X\right),{\mathrm{pv}}_{T+1}^U\left(X\right),{\mathrm{pv}}_{T+1}^D\left(X\right)\}\→$ $S=\{{\mathrm{Trip}}_T^D\left(X\right),{\mathrm{Trip}}_T^U\left(X\right)\},$ Describe external environment condition change and passenger flow or running time fluctuation between probability causal relation and then portray state transfer relationship between the use car demand of external environment condition random disturbance and circuit and available vehicle number $p(S/Z)=\frac{p\left(S\right)\×p\left(Z\right|S)}{p\left(Z\right)}.$

4. the transit scheduling dynamic creation method based on Bayesian network model according to claim 1, it is characterized in that: described step (3) utilizes the real data of intelligent public transportation dispatching platform, total K website and M car, when traffic events being detected, according to public transport environment dynamic change model, the passenger flow arriving k website at time t in conjunction with every i car arrives number and this vehicle arrives the running time at first and last station predict their wave time pt _t(X) and change passenger flow pf _t(X), accordingly according to transit scheduling generation model, gather circuit up-downgoing and use car demand in certain T period ${\mathrm{pr}}_T^U\left(X\right)=\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\underset{i=1}{\overset{M}{\mathrm{\Σ}}}[p_{t-a{t}_t^i-{\mathrm{pt}}_t\left(X\right)}^k+{\mathrm{pf}}_{t-{\mathrm{at}}_t^i-{\mathrm{pt}}_t\left(X\right)}\left(X\right)]{,\mathrm{pr}}_T^D\left(X\right)=\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\underset{i=1}{\overset{M}{\mathrm{\Σ}}}[p_{t-a{t}_t^i-{\mathrm{pt}}_t\left(X\right)}^k+{\mathrm{pf}}_{t-{\mathrm{at}}_t^i-{\mathrm{pt}}_t\left(X\right)}\left(X\right)],$ Available vehicle number ${\mathrm{pv}}_T^U\left(X\right)=\underset{i=1}{\overset{M}{\mathrm{\Σ}}}f({\mathrm{at}}_t^i+{\mathrm{pt}}_t\left(X\right)),{\mathrm{pv}}_T^D\left(X\right)=\underset{i=1}{\overset{M}{\mathrm{\Σ}}}f({\mathrm{at}}_t^i+{\mathrm{pt}}_t\left(X\right))$ And their probability, for transit scheduling establishment provides data supporting.

5. the transit scheduling dynamic creation method based on Bayesian network model according to claim 1, it is characterized in that: described step (4) is on the data basis of step (3), be target around in time evacuating passenger, even equilibrium is adopted to dispatch a car strategy, consider single scheduling method, meet capacity consistency c, the degree of crowding γ ∈ [γ of bus _min, γ _max], on government departure interval F basis, determine up-downgoing departure frequency scope with and their probability, generate possible timetable scheme.

6. the transit scheduling dynamic creation method based on Bayesian network model according to claim 1, it is characterized in that: described step (5) is on the timetable scheme basis of step (4), consider unit mileage running cost l, calculate total passenger's Waiting time of often kind of scheme $\underset{k=1}{\overset{K}{\mathrm{\Σ}}}[\frac{T\·\underset{T}{\∫}{p}_t^k\mathrm{dt}}{\mathrm{Tri}{p}_T^U\left(X\right)}+\frac{T\·\underset{T}{\∫}{p}_t^k\mathrm{dt}}{{\mathrm{Trip}}_T^D\left(X\right)}],$ Website is detained situation $\underset{k=1}{\overset{K}{\mathrm{\Σ}}}[{\mathrm{Trip}}_T^U\left(X\right)\·\mathrm{\γ}\·c-\underset{T}{\∫}{p}_t^k\mathrm{dt}],$ Operation cost indices, and analyze each timetable scheme whether be suitable for traffic environment change, for public traffic management department select the best time show scheme decision support is provided.