Road network global transport capacity risk optimization method based on sensitivity analysis

Abstract

The invention relates to a road network global transport capacity risk optimization method based on sensitivity analysis. The method comprises the following steps: predicting a global transport capacity risk prediction mean value of a road network at a next moment according to a newest passenger flow feature set and a model for a target moment; if the decline amount of the global transport capacity risk prediction mean value is smaller than the target decline amount, sensitivity analysis is carried out according to the newest passenger flow feature set and network parameters of the model to obtain the sensitivity of each passenger flow feature; reducing the target passenger flow volume from the passenger flow volume corresponding to the target passenger flow feature with the highest sensitivity in the plurality of passenger flow features, performing set updating based on the target passenger flow volume, and then executing a global transport capacity risk prediction step and subsequent steps; and outputting each passenger flow feature in the latest passenger flow feature set and the total passenger reduction amount corresponding to each passenger flow feature as a risk optimization result under the condition that the reduction amount of the global transport capacity risk prediction mean value is greater than or equal to the target reduction amount. The optimization speed is high, the time is short, the efficiency is high and the effect is good.

Description

Technical Field

[0001] The present disclosure relates to the field of rail transportation, and in particular to a method for optimizing the global transportation capacity risk of a road network based on sensitivity analysis. Background Art

[0002] Rail transit, with its advantages of large capacity, punctuality, safety, environmental friendliness, and low cost, plays an increasingly important role in modern urban transportation, becoming the backbone and crucial support for modern urban transportation. While rail transit offers greater safety than conventional road transport, due to the large scale of the rail transit network, the heavy transportation workload, and the close coupling between lines, any failure or safety incident can have a significant impact on urban transportation. Therefore, how to effectively assess and predict the overall risk of the rail transit network and thereby reduce this overall risk and its impact on the network's transport capacity has always been a key research issue in the rail transit field. How to predict rail transit network risk while minimizing overall capacity risk based on risk prediction is a technical challenge that needs to be addressed urgently. Summary of the Invention

[0003] In view of this, the present disclosure proposes a method and device for optimizing the global transport capacity risk of a road network based on sensitivity analysis.

[0004] According to one aspect of the present disclosure, a method for optimizing the global transport capacity risk of a road network based on sensitivity analysis is provided, the method comprising: a global transport capacity risk prediction step, a sensitivity analysis step, a set updating step, and a result output step;

[0005] In the global capacity risk prediction step, prediction and calculation are performed based on the latest passenger flow feature set for the target time and the Bayesian network model to obtain a global capacity risk prediction mean value of the rail transit network at the next time, wherein the rail transit network includes at least one line, each line includes at least two stations, and each two stations form an operating interval, and the passenger flow feature set includes multiple passenger flow features of the rail transit network;

[0006] In the sensitivity analysis step, when the decrease in the global capacity risk prediction mean value compared to the initial global capacity risk prediction mean value is less than the target decrease, a sensitivity analysis is performed based on the latest passenger flow feature set and the network parameters of the Bayesian network model to obtain the sensitivity of the global capacity risk prediction mean value to each of the passenger flow features;

[0007] In the set updating step, the passenger flow corresponding to the target passenger flow feature with the highest sensitivity among the multiple passenger flow features is reduced by the target passenger flow, and the set is updated based on the target passenger flow to obtain the latest passenger flow feature set, and the global transport capacity risk prediction step and subsequent steps are executed;

[0008] In the result output step, when the decrease in the global capacity risk prediction mean compared to the initial global capacity risk prediction mean is greater than or equal to the target decrease, each passenger flow feature in the latest passenger flow feature set and the total passenger reduction corresponding to each passenger flow feature are output as risk optimization results.

[0009] First, based on a Bayesian network model for capacity risk prediction, a sensitivity analysis method was employed, combining network centrality to rank the importance of network passenger flow characteristics, and then a gradient descent design approach was employed. With total passenger flow control as the optimization objective, the overall capacity risk was reduced by the required percentage. This optimization approach was fast, time-efficient, and prioritized passenger flow at key stations or sections, which was more effective in reducing overall risk than restricting passenger flow across the entire network.

[0010] In a possible implementation, the global capacity risk prediction step further includes:

[0011] Determine the expected global risk reduction ratio for the next moment after the target moment;

[0012] The target reduction amount is determined based on the expected global risk reduction ratio and the initial global capacity risk prediction mean.

[0013] In a possible implementation, the passenger flow corresponding to the target passenger flow feature with the highest sensitivity among the multiple passenger flow features is reduced by the target passenger flow, and a set update is performed based on the target passenger flow to obtain the latest passenger flow feature set, including:

[0014] According to the sensitivity of each passenger flow feature, the plurality of passenger flow features are sorted in descending order of sensitivity;

[0015] Determine the passenger flow feature ranked first among the multiple passenger flow features as the target passenger flow feature, and reduce the passenger flow corresponding to the target passenger flow feature by the target passenger flow;

[0016] The target passenger flow characteristics are adjusted based on the target passenger flow to obtain the latest target passenger flow characteristics, and the target passenger flow characteristics in the passenger flow characteristic set are replaced with the latest target passenger flow characteristics to obtain the latest passenger flow characteristic set.

[0017] In a possible implementation, the global capacity risk prediction step further includes:

[0018] Obtain a set of passenger flow features for the target time;

[0019] Predictions and calculations are performed based on the passenger flow feature set and the Bayesian network model to obtain an initial global capacity risk prediction mean value of the rail transit network at the next moment.

[0020] In a possible implementation, the multiple passenger flow characteristics include the station passenger flow saturation of each station in the rail transit network, the station stranded passenger evacuation time of each station, and the interval passenger flow saturation of each operating interval.

[0021] In one possible implementation, the Bayesian network model includes multiple layers, which perform predictions and calculations based on the latest passenger flow feature set at the target time and the Bayesian network model to obtain the global capacity risk prediction mean of the rail transit network at the next time, including:

[0022] Inputting the passenger flow saturation of each station and the evacuation time of stranded passengers at the station into the first layer of the Bayesian network model to predict the station capacity risk of each station;

[0023] Inputting the passenger flow saturation of each operating interval into the second layer of the Bayesian network model to predict the interval capacity risk of each operating interval;

[0024] Inputting the station capacity risks of multiple stations on the same line and the interval capacity risks of the operating intervals between adjacent stations into the third layer of the Bayesian network model to predict the line capacity risks of each line;

[0025] Inputting the capacity risk of each line into the fourth layer of the Bayesian network model to predict the global capacity risk distribution value of the rail transit network;

[0026] The mean of the global transport capacity risk distribution values ​​is used as the global transport capacity risk prediction mean.

[0027] In a possible implementation, the target passenger flow includes one of a station passenger flow control amount, a section passenger flow control amount, and a station stranded passenger control amount;

[0028] In the set updating step, the target passenger flow is less than or equal to the corresponding initial passenger flow, where the initial passenger flow is the passenger flow corresponding to the passenger flow feature in the passenger flow feature set used to calculate the initial global transport capacity risk prediction mean.

[0029] According to another aspect of the present disclosure, a device for optimizing road network global transport capacity risk based on sensitivity analysis is provided, the device comprising:

[0030] The global capacity risk prediction module is used to predict and calculate based on the latest passenger flow feature set and Bayesian network model for the target moment to obtain the global capacity risk prediction mean of the rail transit network at the next moment. The rail transit network includes at least one line, each line includes at least two stations, and there is an operating interval between every two stations. The passenger flow feature set includes multiple passenger flow features of the rail transit network, and the multiple passenger flow features include the station passenger flow saturation of each station in the rail transit network, the station stranded passenger evacuation time of each station and the interval passenger flow saturation of each operating interval; the global capacity risk prediction module is also used to send the calculated global capacity risk prediction mean to the sensitivity analysis module and the result output module.

[0031] a sensitivity analysis module configured to, when a decrease in the global capacity risk prediction mean value compared to the initial global capacity risk prediction mean value is less than a target decrease, perform a sensitivity analysis based on the latest passenger flow feature set and the network parameters of the Bayesian network model to obtain a sensitivity of the global capacity risk prediction mean value to each of the passenger flow features;

[0032] a set updating module, configured to reduce the passenger flow corresponding to the target passenger flow feature with the highest sensitivity among the multiple passenger flow features by the target passenger flow, and perform set updating based on the target passenger flow to obtain a latest passenger flow feature set, and send the updated latest passenger flow feature set to the global transport capacity risk prediction module and the result output module;

[0033] The result output module is used to output each passenger flow feature in the latest passenger flow feature set and the total passenger reduction corresponding to each passenger flow feature as a risk optimization result when the decrease in the global capacity risk prediction mean compared to the initial global capacity risk prediction mean is greater than or equal to the target decrease.

[0034] In one possible implementation, the global capacity risk prediction module is further used to determine an expected global risk reduction ratio for the next moment after the target moment; and determine a target reduction amount based on the expected global risk reduction ratio and the initial global capacity risk prediction mean.

[0035] In a possible implementation, the passenger flow corresponding to the target passenger flow feature with the highest sensitivity among the multiple passenger flow features is reduced by the target passenger flow, and a set update is performed based on the target passenger flow to obtain the latest passenger flow feature set, including:

[0036] According to the sensitivity of each passenger flow feature, the plurality of passenger flow features are sorted in descending order of sensitivity;

[0037] Determine the passenger flow feature ranked first among the multiple passenger flow features as the target passenger flow feature, and reduce the passenger flow corresponding to the target passenger flow feature by the target passenger flow;

[0038] The target passenger flow characteristics are adjusted based on the target passenger flow to obtain the latest target passenger flow characteristics, and the target passenger flow characteristics in the passenger flow characteristic set are replaced with the latest target passenger flow characteristics to obtain the latest passenger flow characteristic set.

[0039] In one possible implementation, the global capacity risk prediction module is also used to obtain a passenger flow feature set for the target moment; and to perform predictions and calculations based on the passenger flow feature set and the Bayesian network model to obtain an initial global capacity risk prediction mean value for the rail transit network at the next moment.

[0040] In a possible implementation, the multiple passenger flow characteristics include the station passenger flow saturation of each station in the rail transit network, the station stranded passenger evacuation time of each station, and the interval passenger flow saturation of each operating interval.

[0041] In one possible implementation, the Bayesian network model includes multiple layers, which perform predictions and calculations based on the latest passenger flow feature set at the target time and the Bayesian network model to obtain the global capacity risk prediction mean of the rail transit network at the next time, including:

[0042] Inputting the passenger flow saturation of each station and the evacuation time of stranded passengers at the station into the first layer of the Bayesian network model to predict the station capacity risk of each station;

[0043] Inputting the passenger flow saturation of each operating interval into the second layer of the Bayesian network model to predict the interval capacity risk of each operating interval;

[0044] Inputting the station capacity risks of multiple stations on the same line and the interval capacity risks of the operating intervals between adjacent stations into the third layer of the Bayesian network model to predict the line capacity risks of each line;

[0045] Inputting the capacity risk of each line into the fourth layer of the Bayesian network model to predict the global capacity risk distribution value of the rail transit network;

[0046] The mean of the global transport capacity risk distribution values ​​is used as the global transport capacity risk prediction mean.

[0047] In a possible implementation, the target passenger flow includes one of a station passenger flow control amount, a section passenger flow control amount, and a station stranded passenger control amount;

[0048] The target passenger flow is less than or equal to the corresponding initial passenger flow, and the initial passenger flow is the passenger flow of the corresponding passenger flow feature in the passenger flow feature set used to calculate the initial global transport capacity risk prediction mean.

[0049] According to another aspect of the present disclosure, a road network global capacity risk optimization device based on sensitivity analysis is provided, comprising: a processor; a memory for storing processor executable instructions; wherein the processor is configured to implement the above method when executing the instructions stored in the memory.

[0050] According to another aspect of the present disclosure, a non-volatile computer-readable storage medium is provided, on which computer program instructions are stored, wherein the computer program instructions implement the above method when executed by a processor.

[0051] According to another aspect of the present disclosure, a computer program product is provided, including a computer-readable code, or a non-volatile computer-readable storage medium carrying the computer-readable code. When the computer-readable code runs in a processor of an electronic device, the processor in the electronic device executes the above method.

[0052] According to the method and device for optimizing the global transport capacity risk of a road network based on sensitivity analysis provided by an embodiment of the present disclosure, predictions and calculations are performed based on the latest passenger flow feature set and the Bayesian network model for the target moment to obtain the global transport capacity risk prediction mean value of the rail transit network at the next moment, wherein the rail transit network includes at least one line, each line includes at least two stations, and there is an operating interval between each two stations, and the passenger flow feature set includes multiple passenger flow features of the rail transit network; when the decrease in the global transport capacity risk prediction mean value compared to the initial global transport capacity risk prediction mean value is less than the target decrease value, according to the latest passenger flow feature set and A sensitivity analysis is performed on the network parameters of the Bayesian network model to determine the sensitivity of the global capacity risk prediction mean to each of the passenger flow characteristics. The passenger flow corresponding to the most sensitive target passenger flow characteristic among the multiple passenger flow characteristics is reduced by a target passenger flow, and a set of updated passenger flow characteristics is generated based on the target passenger flow to obtain a latest passenger flow characteristic set. The global capacity risk prediction step and subsequent steps are then executed. If the decrease in the global capacity risk prediction mean compared to the initial global capacity risk prediction mean is greater than or equal to the target decrease, each passenger flow characteristic in the latest passenger flow characteristic set and the total passenger reduction corresponding to each passenger flow characteristic are output as the risk optimization result. Based on the Bayesian network model for capacity risk prediction, a sensitivity analysis method is employed, combining network centrality to rank the importance of network passenger flow characteristics, and then gradient descent is used for design. With the total passenger flow control quantity as the optimization objective, the global capacity risk is reduced by the required proportion. The optimization is fast, time-efficient, and prioritizes passenger flow control at important stations or sections, which is more conducive to reducing global risk than restricting passenger flow across the entire network.

[0053] Further features and aspects of the present disclosure will become apparent from the following detailed description of exemplary embodiments with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate exemplary embodiments, features, and aspects of the disclosure and, together with the description, serve to explain the principles of the disclosure.

[0055] Figure 1 A flowchart of a method for optimizing the global transport capacity risk of a road network based on sensitivity analysis according to an embodiment of the present disclosure is shown.

[0056] Figure 2 A flow chart of a method for optimizing the global transport capacity risk of a road network based on sensitivity analysis according to an embodiment of the present disclosure is shown.

[0057] Figure 3 A block diagram of a road network global transport capacity risk optimization device based on sensitivity analysis according to an embodiment of the present disclosure is shown.

[0058] Figure 4 1 is a block diagram of an apparatus 1900 for optimizing global transport capacity risk of a road network based on sensitivity analysis according to an exemplary embodiment. DETAILED DESCRIPTION

[0059] Various exemplary embodiments, features, and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise indicated.

[0060] The word “exemplary” is used exclusively herein to mean “serving as an example, example, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.

[0061] In addition, numerous specific details are provided in the following detailed description to better illustrate the present disclosure. Those skilled in the art will appreciate that the present disclosure can be practiced without certain specific details. In some instances, methods, means, components, and circuits well known to those skilled in the art are not described in detail in order to highlight the main points of the present disclosure.

[0062] Sensitivity analysis generally studies the impact of model parameters on output. There is also sensitivity analysis for input variables, which studies the impact of changes in input variables on output uncertainty, thereby ranking the importance of parameters or variables. In the regional rail transit capacity risk prediction scenario, a corresponding model can be established (such as the Bayesian network model described in this article). The input variables of this model are the passenger flow characteristics of the rail transit network (also referred to as the road network in this article). Different input variables have different effects on the global capacity risk prediction value output by the model. Therefore, it is necessary to reversely infer the model through sensitivity analysis methods to find the characteristics that have the greatest impact on the global capacity risk of the road network, so as to grasp the weak links of the road network and propose a set of reasonable optimization plans to provide decision support for the subsequent specific scheduling optimization measures.

[0063] Gómez-Villega pointed out that sensitivity analysis is a method for studying the relationship between network inputs and the conditional distribution of the target variable. The inputs can be either parameters in the conditional distribution or the actual values ​​of the observed variables. Laskey was the first to study the complexity of sensitivity analysis in Bayesian networks. Castillo conducted extensive research on sensitivity analysis of discrete and continuous Bayesian networks, using partial derivatives in Gaussian Bayesian networks to calculate the sensitivity of the marginal or conditional probability of the target variable to the values ​​of parameters or evidence variables.

[0064] For continuous Bayesian networks, represented by Gaussian Bayesian networks, according to the sensitivity analysis method proposed by Castillo, given the evidence variable X E =e, target variable X i The probability distribution function can be expressed by the following formula 1:

[0065]

[0066] Where a represents the target variable X i The corresponding setting value.

[0067] Sensitivity analysis is to calculate the partial derivative of the above probability distribution function with respect to the specified parameter θ or evidence variable e, that is, using the following formula 2 or formula 3:

[0068]

[0069]

[0070] Among them, μ(θ;e) and σ(θ;e) are the given evidence variables X E The target variable X i In specific applications, the objects to be calculated for partial derivatives may be the mean, variance, and composite functions of the conditional distribution of the target variable.

[0071] The disclosed embodiment is based on the Castillo sensitivity analysis method, and performs sensitivity analysis on the Bayesian network model for capacity risk prediction (such as the Gaussian Bayesian network model and the Bayesian network model with Gaussian mixture model described in this article). After obtaining the passenger flow of each input node (i.e., station, operating interval (also referred to as interval in this article)) through passenger flow monitoring, that is, the value of the evidence variable, and thus calculating the passenger flow characteristics of the road network "station passenger flow saturation, interval passenger flow saturation, and station stranded passenger evacuation time", belief propagation is used to predict capacity risks at all levels. Now it is necessary to evaluate the sensitivity of the global capacity risk of the road network to the passenger flow of each node under the given node passenger flow.

[0072] For each passenger flow characteristic of the railway network, IP is defined as a single-point sensitivity index, reflecting the degree to which changes in passenger flow, such as station passenger flow saturation, section passenger flow saturation, and waiting time for stranded passengers at stations, affect changes in global capacity risk, as shown in Formulas 4-6. Next, sensitivity analysis is derived for a Gaussian Bayesian network and a Bayesian network with a Gaussian mixture model. For ease of illustration, Table 1 lists the variables involved in the derivation process and relevant to the railway network capacity risk assessment system.

[0073] Table 1 Variable comparison table

[0074] symbol meaning RN(t) Global capacity risk RL(t) Line capacity risk RS(t) Station capacity risk RI(t) Inter-regional capacity risk <![CDATA[R sat (t)]]> Risk of station passenger flow saturation <![CDATA[R wait (t)]]> Risk of passengers being stranded at stations PS(t) Station passenger flow PW(t) Number of stranded passengers at the station PI(t) Passenger flow in the interval

[0075]

[0076]

[0077]

[0078] Among them, IP sat (t) represents the single-point sensitivity index of the station passenger flow saturation risk, IP sec The single point sensitivity index of the interval passenger flow, IP wait (t) represents the single-point sensitivity index of the passenger detention risk at the station.

[0079] (1) Sensitivity analysis of Gaussian Bayesian networks

[0080] According to the Gaussian Bayesian network conditional probability mean estimation expression, it is as follows:

[0081] μ Y|X =β0+β1X1+β2X2+…β k X k Formula 7

[0082] Among them, μ Y|X represents the conditional mean of the random variable Y under the condition of the random variable X, β0, β1…β k Indicates parameters. X1, X2…X k Represent variables respectively.

[0083] The global capacity risk prediction mean Satisfies the following formula 8:

[0084]

[0085] in, are the parameters in the global capacity risk prediction Bayesian network, is the predicted mean value of the transport capacity risk of the kth route. Similarly, the following expressions for the corresponding route transport capacity risk and single-point transport capacity risk can be written:

[0086]

[0087]

[0088]

[0089] in, Indicates station S i The predicted value of passenger flow saturation risk, Indicates station S i The risk of stranded passengers Indicates interval I j risk of passenger flow saturation. represents the predicted capacity risk value of interval n. i, j are the numbers of the corresponding stations / intervals in the road network. Represents calculation The parameters of the Bayesian network used in the process. Indicates the calculation station S i The average value of capacity risk prediction The parameters of the Bayesian network used in the process. Indicates the interval capacity risk RI of interval j j (t) Parameters of the Bayesian network used in the process.

[0090] According to the conditional independence of the Bayesian network, it can be calculated by the chain rule:

[0091]

[0092] Since Formula 12 is only related to the training set data, it is not related to the sample of risk to be optimized. If the existence of transfer nodes is not considered, for station S i , we can get the sensitivity of global transport capacity risk to station passenger flow saturation risk passenger flow PS1(t) The formula is as follows:

[0093]

[0094]

[0095] Among them, w1 i (t) represents the passenger flow saturation risk consequence of station i, PS i (t) represents the passenger flow of station i, f(x) is a sigmoid function, and its expression is From the above formula, we can see that the sensitivity index is not only related to the conditional probability coefficient of the Bayesian network for the station passenger flow saturation risk, but also to the It is related to the current passenger flow saturation and risk consequences w1 i (t) and station S i Station passenger capacity CS i (t) is also related, so when reducing the same passenger flow, the one that reduces the global risk the most is not necessarily The biggest passenger flow feature.

[0096] The single-point sensitivity index can be rewritten as follows:

[0097]

[0098] According to the definition of the f(x) function, the capacity risk f(x) of any given single point is monotonically increasing for the passenger flow characteristics corresponding to the passenger flow, and satisfies f(x)∈(0,1). When the non-negative least squares method is used for parameter learning, the coefficient If is non-negative, then under the Gaussian Bayesian network model, the partial derivative of the global transport capacity risk of the road network with respect to the single-point passenger flow is always greater than zero.

[0099] Considering that S1 is a transfer station, let's assume that S1 is a transfer station between Line 1 and Line 2. Then the sensitivity index of the passenger flow saturation characteristics of station S1 is expressed by formula 15:

[0100]

[0101] This means that transfer stations may be more sensitive to global capacity risk.

[0102] In addition, due to the conditional independence assumption of the linear Gaussian Bayesian network, even if the passenger flow of a station on the same line changes, although it will cause changes in the transport capacity risk and sensitivity index of that node and even the entire line, it will not affect the sensitivity index of the corresponding passenger flow characteristics of another station. This property will have an impact on the subsequent optimization process.

[0103] (2) Sensitivity analysis of mixed Gaussian Bayesian networks

[0104] Based on the parameter learning process of the Bayesian network with a mixture Gaussian model, it can be seen that when using the Bayesian network with a mixture Gaussian model for belief propagation, the prediction of global capacity risk can be written as the following formula 16:

[0105]

[0106] in, μ lRN(t)|RL(t) It is the parameter obtained by the global capacity risk prediction Bayesian network based on the EM algorithm.

[0107] The purpose of the sensitivity analysis of the Bayesian network for capacity risk prediction is to find the gradient of the predicted mean of the global capacity risk variable with respect to the passenger flow corresponding to each input variable, see Formula 4-Formula 6. Since the conditional probability of each layer of the Bayesian network is a mixed Gaussian model, the partial derivative of the global capacity risk prediction value with respect to passenger flow requires the differentiation of multiple layers of mixed Gaussians. At this time, the process of calculating Formula 4-Formula 6 is very complicated. The method of obtaining the gradient using the neural network backpropagation algorithm can be used as a reference, that is, using an algorithm similar to backpropagation gradient descent starting from the global capacity risk output layer, and reversely calculating the expected gradient of the global capacity risk layer by layer according to the chain rule, and storing the calculated gradient for each layer. Therefore, it is only necessary to give the gradient calculation expression of one layer as shown in Formula 17, and the gradient calculation can be realized layer by layer. The calculation method of the single-layer gradient is given below, and the goal is to obtain the following Formula 17:

[0108]

[0109] For the global capacity risk prediction Bayesian network, Equivalent to the average value of the global transport capacity risk prediction of the road network, This is equivalent to the average predicted capacity risk of a certain route. First, based on the results of the EM algorithm, we have the following formula 18:

[0110]

[0111] Where M represents the global transport capacity risk prediction value of the road network, which includes M values, l∈[1,M], j∈[1,M]. l 、μ lS 、μ lR ,∑ lRS ,∑ lSS is the parameter for the lth global transport capacity risk prediction value of the road network, α j 、μ jS ,∑ jSS is the parameter for the jth global capacity risk prediction value of the road network. These are all known quantities for the road network capacity risk prediction and are only relevant to the training set data. Formula 18 can be expanded using the chain rule.

[0112] For linear Gaussian Bayesian networks, due to the properties of linear Gaussian models, given the child nodes, the parent nodes are conditionally independent. S =[x S1 ,x S2 ,…x Sk ] T , there is x S1 ,x S2 ,…x SkHowever, for the Gaussian mixture model, the relationship between the parent nodes is more complex, making it difficult to calculate the gradient through analytical methods. Therefore, the automatic differentiation tool in Python can be used to calculate the gradient and obtain the partial derivatives of the global capacity risk with respect to each input variable passenger flow.

[0113] By calculating the partial derivative of the predicted value of each child node in the Gaussian mixture model with respect to the parent node value, the importance ranking of each input variable can be determined. Controlling passenger flow at a single point is equivalent to constructing a new input variable. Then, using belief propagation to update the node predicted value, the optimized target value can be evaluated. For the Gaussian mixture model, when performing a prediction task, each passenger flow feature is known as an evidence variable, while the conditional probabilities of other nodes with respect to their parent nodes follow a mixed Gaussian distribution. Based on the gradient, gradient descent can be used to further optimize the passenger flow at each input node.

[0114] In rail transit networks, stations are often regarded as basic nodes. Measuring the importance of stations is crucial for optimizing transportation networks, improving service quality, and ensuring safety. In graph theory, centrality is used to measure the importance of network nodes. Centrality can reflect the importance of a station in a rail transit network. Stations with high centrality (also referred to as sites in this article) usually have greater passenger flow and higher risk impact levels in the network. The centrality involved in the network in this disclosed embodiment includes the following:

[0115] Degree Centrality is used to characterize the degree to which a node in an undirected graph is connected to other nodes. It is defined as the number of connections of a node, that is, the number of edges directly connected to the node. D (N i ) represents the number of connections between a station i and its surrounding stations. Stations with high connection centrality usually carry more passenger traffic and are often located at the hub of the rail transit network. Therefore, they have a greater impact on the overall operational risk of the network. As shown in Formula 19, if station j is connected to station i, then x ij is 1 if the value is set, otherwise it is 0.

[0116]

[0117] Among them, C D (N i ) represents the connection centrality of site i in the road network N. g represents the total number of sites in the road network N, i, j∈g.

[0118] Closeness centrality measures the average distance from one station to other stations. The shorter the distance, the higher the closeness centrality. This means that stations with high closeness centrality can connect more quickly to other stations in the network and are often located in the core areas of the rail transit network. Risks occurring at these nodes are more likely to radiate and spread to the surrounding areas. Closeness centrality can be calculated using the following formula 20.

[0119]

[0120] Among them, C C (N i ) represents the distance centrality of site i in the road network N. ij Represents the distance between site i and site j in the road network N.

[0121] Betweenness Centrality measures the number of times a station serves as a transfer station in the shortest path in the rail transit network. jk If (i) is 1, it means that the shortest path from j to k (i.e., from station j to station k) passes through station i. Otherwise, p jk (i) is 0. Stations with high transfer centrality play an important role in connecting different routes or regions in the transportation network. They are often also nodes with highly concentrated passenger flow, significantly impacting the overall capacity risk of the road network. Transfer centrality can be calculated using the following formula 21.

[0122]

[0123] Among them, C B (N i ) represents the transfer centrality of station i in the road network N. jk It represents the number of all routes between station j and station k in the road network N.

[0124] Eigenvector centrality measures the relationship between a station's centrality and the centralities of its directly connected stations. A station's eigenvector centrality is influenced by the centralities of its directly connected stations and the number of lines connected to those stations. Stations with high eigenvector centrality generally have greater influence in the rail transit network. Eigenvector centrality is assessed based on one of the three centrality metrics mentioned above.

[0125] By combining network centrality with sensitivity analysis, the disclosed embodiment can establish a comprehensive importance evaluation index that takes into account the results of the derivation and the structural characteristics of the road network, and is used to determine the order of passenger flow reduction at the nodes of the road network. First, the nodes are preliminarily sorted based on the network centrality, reflecting that the cost of reducing the same amount of passenger flow at different nodes is different. Then, according to the sensitivity analysis method, the impact of each node on the reduction of global capacity risk at the current moment is obtained. The sensitivity of each node is weighted and re-sorted according to the network centrality index, so that the global capacity risk can be minimized to the greatest extent while paying the same cost of passenger flow control.

[0126] In this example, transfer centrality is selected as the metric for measuring node centrality. Since the input variables in the Bayesian network for capacity risk analysis are station passenger flow saturation, stranded passenger evacuation time, and interval passenger flow saturation, weights must also be assigned to interval characteristic nodes. For operating intervals, the average of the centrality indices of the stations at both ends of the interval is used as the centrality metric for the interval nodes.

[0127] On this basis, the disclosed embodiment provides a method for optimizing the global capacity risk of a road network based on sensitivity analysis. The constructed method for optimizing the global capacity risk of a road network is based on the Bayesian network used for the above-mentioned capacity risk situation deduction. According to the Bayesian network model for capacity risk prediction, a sensitivity analysis method is adopted, and the importance of the passenger flow characteristics of the road network is ranked in combination with the network centrality, and then the design is carried out by reference to gradient descent. Taking the total passenger flow control amount as the optimization target, the global capacity risk is reduced to the required proportion, with fast optimization speed, short time and high efficiency, and priority is given to controlling the passenger flow of important stations or sections, which is more conducive to reducing the global risk than restricting the passenger flow of the entire road network.

[0128] The road network global transport capacity risk optimization method based on sensitivity analysis provided in the embodiment of the present disclosure performs optimization with the following formula 22 as the target. is the importance index corresponding to station i, which is calculated from the connection centrality of the road network. It reflects that different stations have different costs to reduce the same passenger flow. The more important the station, the higher the cost of reducing the same passenger flow. The importance index of the same station corresponding to the station passenger flow saturation characteristics and the station passenger retention characteristics is the same, both of which are calculated based on the transfer centrality.

[0129]

[0130] in, Indicates the importance of station i node. Indicates the importance of node j in interval. ΔPS i (t) represents the passenger flow PS at station i i(t) controlled quantity. ΔPI j (t) represents the passenger flow PI of interval j j (t) Control quantity. ΔPW i (t) represents the stranded passengers PW at station i k (t) is the control quantity. L S , L I , L W These are the node lists corresponding to the above three types of passenger flows that need to be controlled.

[0131] Then, in the global transport capacity risk optimization of the road network based on sensitivity analysis, the decision variable is the passenger flow control quantity of each input variable of the transport capacity risk prediction network (i.e., Bayesian network model), i.e., ΔPS i (t), ΔPI j (t), ΔPW i (t). Based on these decision variables, the constraints in the optimization process can be determined as follows:

[0132]

[0133] in, It represents the initial global capacity risk prediction mean of the road network at time t before optimization. represents the expected global transport capacity risk forecast mean value of the road network at time t after optimization. q is the expected global risk reduction ratio. i (t), PI j (t), PW i (t) represents the relevant data of the road network at time t before optimization.

[0134] like Figure 1 As shown, the road network global transport capacity risk optimization method based on sensitivity analysis provided by the embodiment of the present disclosure includes a global transport capacity risk prediction step S101, a sensitivity analysis step S102, a set update step S103 and a result output step S104.

[0135] In the global transport capacity risk prediction step S101, prediction and calculation are performed based on the latest passenger flow feature set for the target time and the Bayesian network model to obtain a global transport capacity risk prediction mean value of the rail transit network at the next time, wherein the rail transit network includes at least one line, each line includes at least two stations, and each two stations are separated by an operating interval. The passenger flow feature set includes multiple passenger flow features of the rail transit network, and the multiple passenger flow features include the station passenger flow saturation of each station in the rail transit network, the station stranded passenger evacuation time of each station, and the interval passenger flow saturation of each operating interval;

[0136] In the sensitivity analysis step S102, when the decrease in the global capacity risk prediction mean value compared to the initial global capacity risk prediction mean value is less than the target decrease, a sensitivity analysis is performed based on the latest passenger flow feature set and the network parameters of the Bayesian network model to obtain the sensitivity of the global capacity risk prediction mean value to each of the passenger flow features;

[0137] In the set updating step S103, the passenger flow corresponding to the target passenger flow feature with the highest sensitivity among the multiple passenger flow features is reduced by the target passenger flow, and the set is updated based on the target passenger flow to obtain the latest passenger flow feature set, and the global transport capacity risk prediction step and subsequent steps are executed;

[0138] In the result output step S104, when the decrease in the global capacity risk prediction mean compared to the initial global capacity risk prediction mean is greater than or equal to the target decrease, each passenger flow feature in the latest passenger flow feature set and the total passenger reduction corresponding to each passenger flow feature are output as risk optimization results.

[0139] This method uses a Bayesian network model for capacity risk prediction, employs sensitivity analysis, and combines network centrality to rank the importance of network passenger flow characteristics. This approach then draws on gradient descent for design. With total passenger flow control as the optimization objective, the overall capacity risk is reduced by the required percentage. This approach is fast, time-efficient, and prioritizes passenger flow control at key stations or sections, which is more effective in reducing overall risk than restricting passenger flow across the entire network.

[0140] To further illustrate the road network global transport capacity risk optimization method based on sensitivity analysis provided by the embodiment of the present disclosure (hereinafter referred to as the optimization method), the following is combined with Figure 2 A schematic description is given.

[0141] like Figure 2 As shown, for each use of the optimization method, it is necessary to first determine the input, which includes a set of passenger flow characteristics at a target time t of the rail transit network (also referred to as the network in this article) to be optimized and an expected global risk reduction ratio q for the network. The rail transit network includes at least one line, each line includes at least two stations, and each two stations are separated by an operating interval. The passenger flow characteristic set includes multiple passenger flow characteristics of the rail transit network, and the multiple passenger flow characteristics include the station passenger flow saturation SS of each station i in the rail transit network. i (t), the evacuation time SW of stranded passengers at each station i i (t) and the passenger flow saturation SI of each operating interval j j (t).

[0142] Then, based on the passenger flow feature set at the target time t in the input and the Bayesian network model, prediction and calculation are performed to obtain the initial global capacity risk prediction mean of the rail transit network at the next moment.

[0143] After that, initialization is performed: according to the expected global risk reduction ratio q and the initial global capacity risk prediction mean Calculate the target descent in, Perform the following loop steps:

[0144] Loop step 1: Calculate the latest global capacity risk forecast mean The corresponding drop in, In particular, since the initial global capacity risk prediction mean is just calculated If a new global capacity risk forecast mean value calculation has not yet been performed, middle Therefore, the initial global capacity risk prediction mean The corresponding drop

[0145] Loop Step 2: Determine Is it less than exist In the case of , continue to execute the loop step 4. , execute the result output step.

[0146] Loop step three: Perform sensitivity analysis based on the latest passenger flow feature set and the network parameters of the Bayesian network model to obtain the sensitivity of the global capacity risk prediction mean to each of the passenger flow features. Reduce the passenger flow corresponding to the target passenger flow feature with the highest sensitivity among the multiple passenger flow features by the target passenger flow, and update the set based on the target passenger flow to obtain the latest passenger flow feature set. In some embodiments, the multiple passenger flow features can be first sorted in descending order of sensitivity according to the sensitivity of each passenger flow feature; the passenger flow feature with the highest ranking among the multiple passenger flow features is determined as the target passenger flow feature, and the passenger flow corresponding to the target passenger flow feature is reduced by the target passenger flow; the target passenger flow feature is adjusted based on the target passenger flow to obtain the latest target passenger flow feature, and the target passenger flow feature in the passenger flow feature set is replaced with the latest target passenger flow feature to obtain the latest passenger flow feature set.

[0147] In some embodiments, the target passenger flow may include the station passenger flow control amount ΔPS i (t), interval passenger flow control quantity ΔPI j (t) and the control quantity of passengers stranded at the station ΔPW i(t). The target passenger flow needs to be less than or equal to the corresponding initial passenger flow, which is the passenger flow corresponding to the passenger flow characteristics in the passenger flow feature set used to calculate the initial global transport capacity risk prediction mean. This ensures that the target passenger flow meets the following constraints:

[0148]

[0149] Among them, PS i (t), PI j (t), PW i (t) respectively represent the initial passenger flow of station i, the initial passenger flow of section j, and the number of stranded passengers at station i, corresponding to the passenger flow characteristics in the passenger flow feature set used for the initial global transport capacity risk prediction mean of the road network at time t before optimization.

[0150] Loop step 4: Based on the latest passenger flow feature set and the Bayesian network model, prediction and calculation are performed to obtain the global capacity risk prediction mean of the rail transit network at the next moment. Then continue to execute loop step one.

[0151] Result output step: outputting each passenger flow feature in the latest passenger flow feature set and the total passenger reduction corresponding to each passenger flow feature as a risk optimization result.

[0152] In one possible implementation, the Bayesian network model may include multiple layers, each performing different predictions. Based on the passenger flow feature set (the input unupdated passenger flow feature set or the updated latest passenger flow feature set) and the Bayesian network model, predictions and calculations are performed to obtain the average global capacity risk prediction for the rail transit network at the next moment, which may include:

[0153] Inputting the passenger flow saturation of each station and the evacuation time of stranded passengers at the station into the first layer of the Bayesian network model to predict the station capacity risk of each station;

[0154] Inputting the passenger flow saturation of each operating interval into the second layer of the Bayesian network model to predict the interval capacity risk of each operating interval;

[0155] Inputting the station capacity risks of multiple stations on the same line and the interval capacity risks of the operating intervals between adjacent stations into the third layer of the Bayesian network model to predict the line capacity risks of each line;

[0156] Inputting the capacity risk of each line into the fourth layer of the Bayesian network model to predict the global capacity risk distribution value of the rail transit network;

[0157] The mean of the global transport capacity risk distribution values ​​is used as the global transport capacity risk prediction mean.

[0158] The sensitivity analysis-based global network capacity risk optimization method provided by the disclosed embodiments utilizes a Bayesian network model for capacity risk prediction, employs sensitivity analysis, and combines network centrality to rank the importance of network passenger flow features, thereby leveraging gradient descent for design. With the total passenger flow control volume as the optimization objective, the method aims to reduce global capacity risk by the required percentage. This method offers rapid, efficient, and time-efficient optimization, prioritizes passenger flow at key stations or sections, and is more conducive to reducing global risk than restricting passenger flow across the entire network.

[0159] like Figure 3 As shown, the embodiment of the present disclosure also provides a road network global transport capacity risk optimization device based on sensitivity analysis, which includes: a global transport capacity risk prediction module 31, a sensitivity analysis module 32, a set update module 33 and a result output module 34. The working process of each module is as follows:

[0160] The global capacity risk prediction module 31 is used to predict and calculate based on the latest passenger flow feature set and the Bayesian network model for the target moment to obtain the global capacity risk prediction mean of the rail transit network at the next moment, wherein the rail transit network includes at least one line, each line includes at least two stations, and there is an operating interval between every two stations. The passenger flow feature set includes multiple passenger flow features of the rail transit network, and the multiple passenger flow features include the station passenger flow saturation of each station in the rail transit network, the station stranded passenger evacuation time of each station, and the interval passenger flow saturation of each operating interval. In some embodiments, the global capacity risk prediction module 31 is also used to send the calculated global capacity risk prediction mean to the sensitivity analysis module 32 and the result output module 34.

[0161] The sensitivity analysis module 32 is used to perform a sensitivity analysis based on the latest passenger flow feature set and the network parameters of the Bayesian network model to obtain the sensitivity of the global capacity risk prediction mean to each of the passenger flow features when the decrease in the global capacity risk prediction mean compared to the initial global capacity risk prediction mean is less than the target decrease.

[0162] The set update module 33 is used to reduce the passenger flow corresponding to the target passenger flow feature with the highest sensitivity among the multiple passenger flow features by the target passenger flow, and perform set update based on the target passenger flow to obtain the latest passenger flow feature set, so as to send the updated latest passenger flow feature set to the global transport capacity risk prediction module 31 and the result output module 34.

[0163] The result output module 34 is used to output each passenger flow feature in the latest passenger flow feature set and the total passenger reduction corresponding to each passenger flow feature as a risk optimization result when the decrease in the global capacity risk prediction mean compared to the initial global capacity risk prediction mean is greater than or equal to the target decrease.

[0164] In one possible implementation, the global capacity risk prediction module is further used to determine an expected global risk reduction ratio for the next moment after the target moment; and determine a target reduction amount based on the expected global risk reduction ratio and the initial global capacity risk prediction mean.

[0165] In a possible implementation, the passenger flow corresponding to the target passenger flow feature with the highest sensitivity among the multiple passenger flow features is reduced by the target passenger flow, and a set update is performed based on the target passenger flow to obtain the latest passenger flow feature set, including:

[0166] According to the sensitivity of each passenger flow feature, the plurality of passenger flow features are sorted in descending order of sensitivity;

[0167] Determine the passenger flow feature ranked first among the multiple passenger flow features as the target passenger flow feature, and reduce the passenger flow corresponding to the target passenger flow feature by the target passenger flow;

[0168] The target passenger flow characteristics are adjusted based on the target passenger flow to obtain the latest target passenger flow characteristics, and the target passenger flow characteristics in the passenger flow characteristic set are replaced with the latest target passenger flow characteristics to obtain the latest passenger flow characteristic set.

[0169] In one possible implementation, the global capacity risk prediction module is also used to obtain a passenger flow feature set for the target moment; and to perform predictions and calculations based on the passenger flow feature set and the Bayesian network model to obtain an initial global capacity risk prediction mean value for the rail transit network at the next moment.

[0170] In a possible implementation, the multiple passenger flow characteristics include the station passenger flow saturation of each station in the rail transit network, the station stranded passenger evacuation time of each station, and the interval passenger flow saturation of each operating interval.

[0171] In one possible implementation, the Bayesian network model includes multiple layers, which perform predictions and calculations based on the latest passenger flow feature set at the target time and the Bayesian network model to obtain the global capacity risk prediction mean of the rail transit network at the next time, including:

[0172] Inputting the passenger flow saturation of each station and the evacuation time of stranded passengers at the station into the first layer of the Bayesian network model to predict the station capacity risk of each station;

[0173] Inputting the passenger flow saturation of each operating interval into the second layer of the Bayesian network model to predict the interval capacity risk of each operating interval;

[0174] Inputting the station capacity risks of multiple stations on the same line and the interval capacity risks of the operating intervals between adjacent stations into the third layer of the Bayesian network model to predict the line capacity risks of each line;

[0175] Inputting the capacity risk of each line into the fourth layer of the Bayesian network model to predict the global capacity risk distribution value of the rail transit network;

[0176] The mean of the global transport capacity risk distribution values ​​is used as the global transport capacity risk prediction mean.

[0177] In a possible implementation, the target passenger flow includes one of a station passenger flow control amount, a section passenger flow control amount, and a station stranded passenger control amount;

[0178] The target passenger flow is less than or equal to the corresponding initial passenger flow, and the initial passenger flow is the passenger flow of the corresponding passenger flow feature in the passenger flow feature set used to calculate the initial global transport capacity risk prediction mean.

[0179] Through the above-mentioned road network global transport capacity risk optimization device based on sensitivity analysis, the global transport capacity risk prediction module is used to predict and calculate according to the latest passenger flow feature set and the Bayesian network model for the target moment, and obtain the global transport capacity risk prediction mean of the rail transit network at the next moment, the rail transit network includes at least one line, each line includes at least two stations, and there is an operating interval between every two stations, and the passenger flow feature set includes multiple passenger flow features of the rail transit network; the sensitivity analysis module is used to, when the decrease in the global transport capacity risk prediction mean compared with the initial global transport capacity risk prediction mean is less than the target decrease, calculate the global transport capacity risk prediction mean according to the latest passenger flow feature set and the Bayesian network model. A sensitivity analysis is performed on the network parameters of the Bayesian network model to determine the sensitivity of the global capacity risk prediction mean to each of the passenger flow characteristics. A set updating module is configured to reduce the passenger flow corresponding to the most sensitive target passenger flow characteristic among the multiple passenger flow characteristics by a target passenger flow, perform a set update based on the target passenger flow to obtain a latest passenger flow characteristic set, and execute the global capacity risk prediction step and subsequent steps. A result output module is configured to output each passenger flow characteristic in the latest passenger flow characteristic set and the total passenger reduction corresponding to each passenger flow characteristic as a risk optimization result, if the decrease in the global capacity risk prediction mean compared to the initial global capacity risk prediction mean is greater than or equal to the target decrease. Based on the Bayesian network model for capacity risk prediction, a sensitivity analysis method is employed, combining network centrality to rank the importance of passenger flow characteristics in the road network, and then a gradient descent design is employed. With the total passenger flow control quantity as the optimization objective, the global capacity risk is reduced to the required ratio. The optimization is fast, time-efficient, and prioritizes passenger flow control at important stations or sections, which is more conducive to reducing global risk than restricting passenger flow across the entire road network.

[0180] In some embodiments, the functions or modules included in the device provided by the embodiments of the present disclosure can be used to execute the method described in the above method embodiments. Its specific implementation and beneficial effects can refer to the description of the above method embodiments. For the sake of brevity, they will not be repeated here.

[0181] It should be noted that while the above embodiments serve as examples for describing a method and apparatus for optimizing global road network capacity risk based on sensitivity analysis, those skilled in the art will appreciate that the present disclosure is not limited thereto. In fact, users can flexibly configure the various steps and modules based on their personal preferences and / or actual application scenarios, as long as they comply with the technical solutions of the present disclosure.

[0182] The present disclosure also provides a computer-readable storage medium having computer program instructions stored thereon, wherein the computer program instructions implement the above method when executed by a processor. The computer-readable storage medium may be a volatile or non-volatile computer-readable storage medium.

[0183] An embodiment of the present disclosure further proposes an electronic device, comprising: a processor; and a memory for storing instructions executable by the processor; wherein the processor is configured to implement the above method when executing the instructions stored in the memory.

[0184] An embodiment of the present disclosure also provides a computer program product, including computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code. When the computer-readable code runs in a processor of an electronic device, the processor in the electronic device executes the above method.

[0185] Figure 4 1 is a block diagram of an apparatus 1900 for optimizing global transport capacity risk of a road network based on sensitivity analysis according to an exemplary embodiment. For example, the apparatus 1900 can be provided as a server or a terminal device. Figure 4 The apparatus 1900 includes a processing component 1922, which further includes one or more processors, and a memory resource represented by a memory 1932 for storing instructions, such as an application, that can be executed by the processing component 1922. The application stored in the memory 1932 may include one or more modules, each corresponding to a set of instructions. In addition, the processing component 1922 is configured to execute the instructions to perform the above-described method.

[0186] The device 1900 may also include a power supply component 1926 configured to perform power management of the device 1900, a wired or wireless network interface 1950 configured to connect the device 1900 to a network, and an input / output interface 1958 (I / O interface). The device 1900 may operate based on an operating system stored in the memory 1932, such as Windows Server 2003. TM , MacOS X TM , Unix TM ,Linux TM , FreeBSD TM or similar.

[0187] In an exemplary embodiment, a non-volatile computer-readable storage medium is also provided, such as a memory 1932 including computer program instructions that can be executed by the processing component 1922 of the apparatus 1900 to perform the above-described method.

[0188] The present disclosure may be a system, method and / or computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0189] A computer-readable storage medium can be a tangible device that can hold and store instructions for use by an instruction execution device. A computer-readable storage medium can be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punch card or a raised structure in a groove on which instructions are stored, and any suitable combination thereof. As used herein, a computer-readable storage medium is not to be construed as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse through a fiber optic cable), or an electrical signal transmitted through an electrical wire.

[0190] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions to be stored in the computer-readable storage medium in each computing / processing device.

[0191] The computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, and conventional procedural programming languages ​​such as "C" language or similar programming languages. Computer-readable program instructions may be executed entirely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., utilizing an Internet service provider to connect via the Internet). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may be personalized by utilizing the state information of the computer-readable program instructions. The electronic circuit may execute the computer-readable program instructions, thereby realizing various aspects of the present disclosure.

[0192] Various aspects of the present disclosure are described herein with reference to flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present disclosure. It should be understood that each block of the flowcharts and / or block diagrams, and combinations of blocks in the flowcharts and / or block diagrams, can be implemented by computer-readable program instructions.

[0193] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processor of the computer or other programmable data processing device, a device is generated that implements the functions / actions specified in one or more blocks in the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium, where these instructions cause the computer, programmable data processing device, and / or other device to operate in a specific manner. Thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing various aspects of the functions / actions specified in one or more blocks in the flowchart and / or block diagram.

[0194] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operational steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / actions specified in one or more blocks in the flowchart and / or block diagram.

[0195] The flow charts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the systems, methods and computer program products according to multiple embodiments of the present disclosure. In this regard, each box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and the part of the module, program segment or instruction contains one or more executable instructions for realizing the prescribed logical function. In some alternative implementations, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the prescribed function or action, or can be implemented by a combination of dedicated hardware and computer instructions.

[0196] While various embodiments of the present disclosure have been described above, the foregoing description is intended to be illustrative, non-exhaustive, and not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, their practical applications, or technological improvements in the marketplace, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method for optimizing global transport capacity risk of a road network based on sensitivity analysis, characterized in that: The method comprises: a global transport capacity risk prediction step, a sensitivity analysis step, a set updating step and a result output step; In the global capacity risk prediction step, prediction and calculation are performed based on the latest passenger flow feature set for the target time and the Bayesian network model to obtain a global capacity risk prediction mean value of the rail transit network at the next time, wherein the rail transit network includes at least one line, each line includes at least two stations, and each two stations form an operating interval, and the passenger flow feature set includes multiple passenger flow features of the rail transit network; In the sensitivity analysis step, when the decrease in the global capacity risk prediction mean value compared to the initial global capacity risk prediction mean value is less than the target decrease, a sensitivity analysis is performed based on the latest passenger flow feature set and the network parameters of the Bayesian network model to obtain the sensitivity of the global capacity risk prediction mean value to each of the passenger flow features; In the set updating step, the passenger flow corresponding to the target passenger flow feature with the highest sensitivity among the multiple passenger flow features is reduced by the target passenger flow, and the set is updated based on the target passenger flow to obtain the latest passenger flow feature set, and the global transport capacity risk prediction step and subsequent steps are executed; In the result output step, when the decrease in the global capacity risk prediction mean compared to the initial global capacity risk prediction mean is greater than or equal to the target decrease, each passenger flow feature in the latest passenger flow feature set and the total passenger reduction corresponding to each passenger flow feature are output as risk optimization results.

2. The method according to claim 1, characterized in that The global capacity risk prediction step also includes: Determine the expected global risk reduction ratio for the next moment after the target moment; The target reduction amount is determined based on the expected global risk reduction ratio and the initial global capacity risk prediction mean.

3. The method according to claim 1, characterized in that The method further comprises: reducing the passenger flow corresponding to the target passenger flow feature with the highest sensitivity among the plurality of passenger flow features by the target passenger flow, and performing a set update based on the target passenger flow to obtain a latest passenger flow feature set, including: According to the sensitivity of each passenger flow feature, the plurality of passenger flow features are sorted in descending order of sensitivity; Determine the passenger flow feature ranked first among the multiple passenger flow features as the target passenger flow feature, and reduce the passenger flow corresponding to the target passenger flow feature by the target passenger flow; The target passenger flow characteristics are adjusted based on the target passenger flow to obtain the latest target passenger flow characteristics, and the target passenger flow characteristics in the passenger flow characteristic set are replaced with the latest target passenger flow characteristics to obtain the latest passenger flow characteristic set.

4. The method according to claim 1, wherein The global capacity risk prediction step also includes: Obtain a set of passenger flow features for the target time; Predictions and calculations are performed based on the passenger flow feature set and the Bayesian network model to obtain an initial global capacity risk prediction mean value of the rail transit network at the next moment.

5. The method according to claim 1, wherein The multiple passenger flow characteristics include the station passenger flow saturation of each station in the rail transit network, the station stranded passenger evacuation time of each station, and the interval passenger flow saturation of each operating interval.

6. The method according to claim 5, characterized in that The Bayesian network model includes multiple layers, and performs predictions and calculations based on the latest passenger flow feature set at the target time and the Bayesian network model to obtain the global capacity risk prediction mean of the rail transit network at the next moment, including: Inputting the passenger flow saturation of each station and the evacuation time of stranded passengers at the station into the first layer of the Bayesian network model to predict the station capacity risk of each station; Inputting the passenger flow saturation of each operating interval into the second layer of the Bayesian network model to predict the interval capacity risk of each operating interval; Inputting the station capacity risks of multiple stations on the same line and the interval capacity risks of the operating intervals between adjacent stations into the third layer of the Bayesian network model to predict the line capacity risks of each line; Inputting the capacity risk of each line into the fourth layer of the Bayesian network model to predict the global capacity risk distribution value of the rail transit network; The mean of the global transport capacity risk distribution values ​​is used as the global transport capacity risk prediction mean.

7. The method according to claim 1, characterized in that The target passenger flow includes one of the station passenger flow control volume, section passenger flow control volume and station stranded passenger control volume; In the set updating step, the target passenger flow is less than or equal to the corresponding initial passenger flow, where the initial passenger flow is the passenger flow corresponding to the passenger flow feature in the passenger flow feature set used to calculate the initial global transport capacity risk prediction mean.

8. A road network global transport capacity risk optimization device based on sensitivity analysis, characterized in that: The device comprises: a global transport capacity risk prediction module, configured to predict and calculate based on the latest passenger flow feature set for the target moment and a Bayesian network model to obtain a global transport capacity risk prediction mean value for the rail transit network at the next moment, wherein the rail transit network includes at least one line, each line includes at least two stations, and each two stations are separated by an operating interval, and the passenger flow feature set includes multiple passenger flow features of the rail transit network, wherein the multiple passenger flow features include the station passenger flow saturation of each station in the rail transit network, the station stranded passenger evacuation time of each station, and the interval passenger flow saturation of each operating interval; a sensitivity analysis module configured to, when a decrease in the global capacity risk prediction mean value compared to the initial global capacity risk prediction mean value is less than a target decrease, perform a sensitivity analysis based on the latest passenger flow feature set and the network parameters of the Bayesian network model to obtain a sensitivity of the global capacity risk prediction mean value to each of the passenger flow features; a set updating module, configured to reduce the passenger flow corresponding to the target passenger flow feature with the highest sensitivity among the multiple passenger flow features by the target passenger flow, and perform set updating based on the target passenger flow to obtain a latest passenger flow feature set, and send the updated latest passenger flow feature set to the global transport capacity risk prediction module and the result output module; The result output module is used to output each passenger flow feature in the latest passenger flow feature set and the total passenger reduction corresponding to each passenger flow feature as a risk optimization result when the decrease in the global capacity risk prediction mean compared to the initial global capacity risk prediction mean is greater than or equal to the target decrease.

9. A road network global transport capacity risk optimization device based on sensitivity analysis, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to implement the method according to any one of claims 1 to 8 when executing the instructions stored in the memory.

10. A non-volatile computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images