Time domain optimization method of urban rail transit based on spatiotemporal distribution of passenger flow

Abstract

A kind of urban rail transit time domain optimization method based on passenger flow space-time distribution deduction, comprising the following steps: OD passenger flow is input, passenger travel network is constructed, dynamic cost-based multi-path search algorithm is used to carry out passenger space-time path deduction, restore passenger travel path, and the space-time distribution state of line passenger flow is counted;With line passenger flow space-time state distribution as input, operation time domain optimization is carried out using ordered clustering method, and the number of time domain optimization is determined;Based on the principle of ordered sample clustering, the urban rail operation time domain optimization framework is constructed, and the two-way operation time domain integration is carried out.The advantages of the present application are that: the time domain optimization framework supports small granularity and multi-direction time domain optimization, introduces the concept of characteristic passenger flow and characteristic direction, can become the link between vehicle flow and driving, realizes accurate and flexible compilation, and realizes cost reduction and efficiency increase of operation through reasonable vehicle flow organization scheme under the premise of guaranteeing service level.

Description

Technical Field

[0001] This invention relates to an optimization method, and more particularly to a time-domain optimization method for urban rail transit based on the spatiotemporal distribution of passenger flow. The invention proposes a new time-domain optimization technology for urban rail transit operation, which, based on the existing line capacity constraints and train operation sequence, highly couples capacity and passenger volume. By applying an operation time-domain division framework with the spatiotemporal distribution of passenger flow as input, it performs bidirectional operation time-domain integration to achieve a reasonable operation time-domain division. Background Technology

[0002] Passenger flow and train operation are the core of urban rail transit operations. Developing operation plans that align with passenger flow characteristics is crucial for providing high-quality transportation services to the public. Since passenger flow changes continuously throughout the day, its characteristics are relatively consistent within a fixed timeframe. Therefore, it is necessary to find the start and end times of a relatively stable passenger flow timeframe through an effective method, thus defining the operational timeframe. A reasonable timeframe division allows operation planners to flexibly formulate train operation plans while meeting passenger demand, achieving the goals of cost reduction and efficiency improvement in urban rail transit operations.

[0003] In traditional urban rail transit operation planning, the daily operating time domain is implicitly divided into several specific time domains, such as early departure, morning peak, off-peak, evening peak, and evening off-peak. Research specifically targeting time domain optimization for urban rail transit is limited. Existing methods partially consider only unidirectional origin-destination (OD) passenger flow, which is not well-suited for lines connecting urban and suburban areas with significant spatial imbalances in passenger flow. In actual operation planning, especially for long-distance lines with extended train turnaround times, existing literature on time domain optimization does not consider the impact of train operation on time domain division.

[0004] In traditional urban rail transit operation planning, the daily operating time domain is by default divided into several specific time domains such as early departure, morning peak, off-peak, evening peak, and evening off-peak. Research specifically targeting urban rail transit time domain optimization is limited. Sun Yan (Determination of Urban Rail Transit Train Operation Scheme, Tongji University Journal) constructed an operation scheme optimization model with the objectives of minimizing the number of operating time domains and minimizing the mismatch between transport capacity and passenger volume. It optimizes the operating time domain based on the minimum difference in the number of different trains within adjacent time periods. Xu Dejie (Optimization of Train Operation Scheme Considering the Operation Ratio, Journal of Transportation Engineering) uses passenger flow entering and exiting stations as input and merges the operating time domains using the passenger flow time domain coefficients during morning and evening peak hours. Xing Xiaoyu (Research on Metro Train Timetable Optimization Based on Passenger Flow Spatiotemporal Characteristics Cluster Analysis, Beijing Jiaotong University Dissertation) uses passenger flow entering and exiting stations as input and employs the k-means clustering method to analyze urban rail transit passenger flow and complete the division of the urban rail transit operating time domain. These methods failed to capture the spatial variation patterns of passenger flow within the operational time domain, and their time domain division accuracy was relatively low, indicating significant limitations in time domain optimization. To further improve the accuracy of time domain division, Zeng Xiaoxu (Application of Ordered Sample Clustering Method in Urban Rail Transit Operation Time Division: Urban Rapid Rail Transit) proposed an operational time domain division method based on unidirectional OD probability matrix and ordered sample clustering, using 20-minute granularity OD passenger flow as input. The method aims to find the optimal operational time domain division method with the minimum cost loss function value as the objective. However, since this method only considers unidirectional OD passenger flow, it lacks adaptability for lines connecting urban and suburban areas with strong passenger flow spatial imbalance. In actual operation planning, especially for long-distance lines with long train turnaround times, existing literature on time domain optimization does not consider the impact of train operation on time domain division. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a time-domain optimization method for urban rail transit based on the spatiotemporal distribution of passenger flow. This technical solution includes the following steps: Step S1: Using OD passenger flow as input, construct a passenger travel network, use a multi-path search algorithm based on dynamic cost to extrapolate passenger spatiotemporal paths, reconstruct passenger travel paths, and statistically analyze the spatiotemporal distribution of passenger flow on the route. Step S2: Using the spatiotemporal distribution of passenger flow as input, perform time-domain optimization using ordered clustering and determine the number of time-domain optimizations. Step S3: Based on the ordered sample clustering results, construct a time-domain optimization framework for urban rail transit operation and perform bidirectional time-domain integration.

[0006] Compared with the prior art, the advantages of the present invention are as follows: the time-domain optimization framework, on the basis of supporting fine-grained and multi-directional time-domain optimization, introduces the concepts of characteristic passenger flow and characteristic direction, which can become the link coupling traffic flow and driving, realize precise and flexible planning, and achieve cost reduction and efficiency improvement in operation through reasonable traffic flow organization scheme while ensuring service level. Attached Figure Description

[0007] Figure 1. Schematic diagram of time-domain partitioning example; Figure 2 shows a schematic diagram of the maximum cross-sectional passenger flow at different times; Figure 3. Schematic diagram of the changing trend of the uplink and downlink sample clustering loss function; Figure 4. Schematic diagram of time-domain partitioning results; Figure 5. Schematic diagram of the initial uplink and downlink time domain partitioning results when k=26; Figure 6. Schematic diagram of the initial time domain partitioning results of uplink and downlink samples when k=26. Detailed Implementation

[0008] To make the objectives, technical solutions, design methods, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention.

[0009] According to an embodiment of the present invention, a time-domain optimization method for urban rail transit based on the extrapolation of passenger flow spatiotemporal distribution is provided.

[0010] This method can be summarized as follows: It utilizes a dynamic cost multi-path search algorithm to statistically analyze the spatiotemporal distribution of passenger flow along railway lines; it combines multi-probability travel path selection to extrapolate the spatiotemporal paths of passenger flow; and based on the principle of ordered sample clustering, it constructs a time-domain optimization framework for urban rail transit operation, performing bidirectional time-domain integration. Specifically, this technology includes the following steps: Step S1: Using OD passenger flow as input, construct a passenger travel network, employ a multi-path search algorithm based on dynamic costs to extrapolate passenger spatiotemporal paths, reconstruct passenger travel routes, and statistically analyze the spatiotemporal distribution of passenger flow on the routes. Based on the urban rail network topology, a passenger travel path search network is constructed, comprising ordinary nodes, transfer nodes, direct arcs, and transfer arcs. To consider the sensitivity of passenger travel to changes in travel time and number of transfers, this paper proposes a method for calculating transfer arc costs based on path travel time and number of transfers, fully describing the impact of transfers on passenger travel. For the set of transfer arcs, The transfer penalty coefficient is the origin of the travel network. The destination is The path, passing through the arc Costs incurred The calculation method is shown in the following formula. This invention uses the A* algorithm for K-shortest path search in a travel network. The A* algorithm combines the characteristics of heuristic and formal methods, and is an effective algorithm for path search in static networks. The evaluation function, as the key to the A* algorithm, is expressed as follows: ,in This represents the distance from the starting point to the current node. Path cost, This represents the distance from the endpoint to the current node. The shortest path cost. (Set) Used to represent the distance from the starting point to the current point. The path is used in the algorithm to avoid searching for a roundabout path.

[0011] The steps for K-short-circuit search based on the passenger travel network are as follows: Step 1: Initialize the road network and establish a directed network for passenger travel; Step 2: Calculate the endpoint using the SPFA algorithm The shortest path to all other points in the network is obtained. .

[0012] set up For nodes The set of predecessor nodes, Recorded as express Arc cost, From the starting point to the node The set of paths To verify the node queue, the specific steps are as follows: ① Initialization, let , , , , ; ② From the queue Extract the first element from the list and use it as the current node. ; ③ Loop through the current node Forward node set Take node ; ④ Calculate the arc cost according to the formula. Determine whether and If so, then update. and ; ⑤ Determine if the update node is in the queue. If it is not in the queue, add it to the queue; ⑥ Determine the set Has the loop ended? If so, remove from the queue. Remove node Otherwise, repeat steps ④-⑤; ⑦ Determine the set Is it If yes, the algorithm ends; otherwise, repeat steps ③-⑥.

[0013] Step 3: Use the A* algorithm to search for K-short circuits between the travel networks.

[0014] set up For nodes The set of subsequent nodes, Recorded as express Arc cost, From the starting point to the node The set of paths Starting point to The The path cost of the shortest circuit. To verify the node queue, The specific steps are as follows: ① Initialization, let , , , ; ② From the queue Extract the first element from the list and use it as the current node. ; ③ Loop through the current node The set of subsequent nodes Take node ; ④ Determine if the node is If the arc cost is calculated according to the formula ,renew , , ; ⑤ Determine if the node is If so, then output the first... The sum of the costs of the arcs traversed along the short circuit path is denoted as: ,make Otherwise Add to queue middle; ⑥ Determine the set Has the loop ended? If so, remove from the queue. Remove node ,queue Sort in ascending order according to the valuation function; otherwise, repeat steps ④-⑤. ⑦ Determine the set Is it or If yes, the algorithm ends; otherwise, repeat steps ③-⑥. Step 4: Based on the correspondence between the travel network and the topology network, reconstruct... Between Short circuit.

[0015] Distributing passenger flow across the entire network based on K-short-circuit search results: Since passengers tend to choose lower-cost routes during their travels, and the cost of the k-th path (starting from o and ending at d) in the K-path search process... The calculations have already taken into account passengers' sensitivity to travel time and the number of transfers. Based on this, this invention uses a Logit model to calculate the probability of path selection, starting from... The destination is Passenger routes The proportion can be expressed in the following form, where This is a non-negative parameter, and its magnitude represents the passenger's familiarity with the route. The larger the value, the greater the probability that the passenger will choose that route.

[0016] Starting point is The destination is Passengers choose the first Passenger flow along the route The calculation formula is as follows: Travel Network Arc Passenger flow The calculation formula is as follows, where Indicates passing through arc The set of paths.

[0017] In actual travel, various factors can influence passenger flow allocation, leading to discrepancies between the actual passenger flow and the actual travel experience. To facilitate rapid passenger flow allocation within acceptable margins, this invention temporarily disregards the impact of train operation on passenger flow allocation. Passenger entry time, transfer travel time, and train stop time are all included in the travel time minutes of the travel network arc. The specific passenger flow allocation algorithm is as follows: Step 1: Initialize the passenger travel network and start passenger flow allocation. Passenger flow allocation end time passenger flow allocation granularity The basic time domain set for passenger flow allocation is .

[0018] Step 2: Calculate the travel network Between Shortest path.

[0019] Step 3: Calculate the selection ratio of each path.

[0020] Step 4: Take For the current time domain, statistical analysis is conducted within each time domain. Passenger flow value .

[0021] Step 5: Take For the present Determine whether If so, then calculate. They are then assigned to the corresponding travel networks.

[0022] Step 6: Determine if completed Loop, if so, then take Backward time domain If it is the current time domain, then repeat step 5.

[0023] Step 7: Calculate the arc flow of the travel network in each time domain. Based on the arc type, the passenger flow at different time sections and the passenger flow for different time transfers can be obtained.

[0024] Step S2: Using the spatiotemporal distribution of passenger flow on the line as input, the ordered clustering method is used to optimize the operation in the time domain, and the number of time domain optimizations is determined.

[0025] Using the principle of ordered sample clustering, it is necessary to define the diameter of the cluster, as follows: Assumption yes An ordered sample, for 3D vector. Let one class be... Includes samples , recorded as Define the diameter of this class as... In the formula This represents the sum of squared deviations of the sample. The mean vector of this class Step 1: Using the characteristic day's OD passenger flow as input, and through the passenger flow spatiotemporal path extrapolation framework, statistically analyze the time-sharing cross-sectional passenger flow of the selected route. The statistical analysis results of the maximum time-sharing cross-sectional passenger flow are as follows: Figure 2 As shown.

[0026] Step 2: Initialize clustering parameters and samples.

[0027] ① Read the topology data of the research line and the passenger flow data of the characteristic day's time section, and set the operation start time as... The operation will end at [time]. Minimum granularity of time domain partitioning Set constraints on the number of time-domain partitions. .

[0028] ②Time Sample In the formula Representing the time domain The passenger flow state vector of the inward-flow section, where To study the first line A cross-section in the time domain Similarly, the upstream time sample can be obtained from the cross-sectional passenger flow. .

[0029] Step 3: Select the optimal number of time domain partitions .

[0030] Introducing a classification loss function, its contents include: use Indicates will The ordered samples are divided into A certain way of classifying, denoted as In the formula, For classification indexing. Classification loss function definition: exist , At a certain time, The smaller the value, the higher the similarity between categories, and the more reasonable the classification.

[0031] The goal of the optimal segmentation scheme is to find the classification method that minimizes the classification loss function among all classification results. .Right now: The formula indicates that The ordered samples are divided into The optimal segmentation scheme for a class is based on the previous an ordered sample Optimal segmentation. Given the optimal number of segments, the optimal segmentation scheme can be obtained based on the above formula. .

[0032] The ordered clustering method is used to solve the classification loss function value under different time domain divisions for uplink and downlink. ,draw Follow Based on the trend chart and the curve's trajectory, determine the optimal number of time-domain partitions. Based on this, the number of time-domain partitions to solve is... The optimal time-domain partitioning scheme, where the optimal downlink partitioning scheme is: The optimal uplink partitioning scheme is The division results are based on... Figure 3 For example, the classification loss function is obtained by combining the examples. ( ) with the number of categories The changing trend, where k takes the value of 26.

[0033] Step S3: Based on the principle of ordered sample clustering, construct a time-domain optimization framework for urban rail transit operation and perform bidirectional time-domain integration. Step 1: Standardize the results of unidirectional time domain partitioning.

[0034] Considering that the capacity allocation in actual operation planning needs to meet the capacity demand of the maximum passenger flow section, and in order to facilitate the layout of train operation lines, the time domain division should match the originating station of the corresponding train class. Based on this, this paper proposes the concept of time domain characteristic passenger flow density to standardize the unidirectional time domain division results. The specific steps are as follows: ① Based on the optimal time-domain partitioning results, find the time-domain maximum passenger flow cross-section, and calculate the time-domain characteristic passenger flow density based on the passenger flow value of the maximum cross-section using the following formula. Where Representing the time domain Characteristic passenger flow density, Representing the time domain Maximum cross-sectional passenger flow Representing the time domain The length, in units of , This indicates the duration of the fundamental time domain.

[0035] ② Using the maximum passenger flow section as the initial section, the time-domain division result is standardized based on the directed interval running time and station stopping time, where... Representing the time domain The start or end time, Representing the time domain The total stopping time of the train from its originating station to the reference section. Representing the time domain The total travel time of a train from its originating station to the reference section. This represents the start and end times of the standardized time domain.

[0036] ③ Based on the passenger flow density in the time domain, determine whether the difference between adjacent time domain intervals is less than the minimum train headway time limit in the time domain. If so, then the adjacent time-domain partitioning results are merged to complete the standardization. In the formula... Representing the time domain Train intervals, , These represent the minimum and maximum headway times for the selected route, respectively. This indicates the train's rated passenger capacity, expressed in person / car. Representing the time domain The expected occupancy rate of the train This indicates the duration of the fundamental time domain.

[0037] Step 2: Integrate the results of bidirectional time domain partitioning.

[0038] For time domains with non-transitional attributes, in order to meet the turnaround needs of trains within the time domain, even if there is a waste of unidirectional time domain capacity, the consistency of the up and down time domain division should be maintained, and the integration of the bidirectional operation time domain division results should be completed based on this.

[0039] Based on the number of sample categories Taking 26 o'clock as an example, the initial uplink and downlink time domain division results are as follows: Figure 5 As shown, the granularity of the time domain division can reach 5 minutes. The uplink time domain segmentation points are concentrated in the afternoon time domain, while the downlink time domain segmentation points are concentrated in the morning time domain. The time domain division result conforms to the spatiotemporal imbalance characteristics of the selected line. To ensure the feasibility of the time domain division result in actual operation, the initial uplink and downlink time domain divisions are integrated according to the urban rail transit operation time domain optimization framework to obtain the number of time domains. The time domain division times are shown in [link to time domain division details]. Figure 6 In this operational time-domain division scheme, the required transport capacity is used as the standard for measuring the stability of passenger flow characteristics in each time domain. The minimum headway difference between each merged time domain is within 20 seconds, indicating stable passenger flow characteristics. The minimum transport capacity demand difference between adjacent time domains exceeds 20 seconds, meeting the basic conditions for adjusting urban rail train headway. For example... Figure 4 As shown, the time-domain optimization results can cover the basic time domain with similar characteristic passenger flows quite well.

[0040] It should be noted that the final time-domain partitioning result includes characteristic direction attributes, used to represent the spatial unevenness of passenger flow distribution within the time domain. For time domains with longer durations, to meet the turnaround requirements of trains within the time domain, the corresponding non-characteristic direction time-domain partitioning should be consistent with the characteristic direction. Shorter duration time domains generally belong to the transition time domain between off-peak and peak hours; when analyzing time-domain route schemes, this type of time domain should be merged with adjacent off-peak time domains. This paper partitions shorter duration time domains separately to highlight the changes in passenger flow between time domains and guide the generation of train operation lines in transition attribute time domains. Considering the differences in arrival and departure routes due to the different locations of different line sections, the non-characteristic direction time-domain partitioning results of transition attribute time domains can be obtained from the initial time-domain partitioning result table.

[0041] Although the steps are described in a specific order above, it does not mean that they must be executed in that specific order. In fact, some of these steps can be executed concurrently or even in a different order, as long as the required functionality can be achieved.

[0042] In one embodiment, the urban rail line needs to have two main lines, each of which can only be used by passengers traveling in one direction. The line has multiple stations, which can be divided into transfer stations and non-transfer stations based on their function in passenger travel. Stations divide the line into several sections, and passengers travel along these sections with the train. Based on the origin-destination (OD) characteristics of passenger travel, passengers can be divided into local line passengers and cross-line passengers. Local line passengers' travel includes entering the station, boarding, alighting, and exiting the station. Cross-line passengers' travel includes entering the station, boarding, alighting, transferring, boarding again, alighting, and exiting the station. Based on network passenger travel behavior analysis, single-line basic time-domain passenger flow distribution statistics are performed. Considering the similarity of passenger flow characteristics across different time domains and the feasibility in actual operation, line time-domain optimization is carried out.

[0043] by Figure 1 Taking the route shown as an example, this route includes 7 stations and 6 sections. The stations are numbered sequentially along the downhill direction as follows: The intervals along the downward direction are as follows: .station and This is a transfer station; passengers whose origin or destination is not on this line can choose the corresponding transfer station to complete their journey based on their travel needs. The line's operating time zone begins at [time]. The end time is , including the basic time domain Characteristic passenger flow Based on the time domain Similarly, the downward passenger flow characteristics Based on the time domain The upward-moving passenger flow. For example... Figure 1As shown, taking the downstream time domain division as an example, adjacent passenger flow with similar characteristics are grouped into one category, and the initial time domain division result is: , , During the operation plan development process, in order to meet the maximum cross-sectional area requirement... Based on the capacity demand, combined with the location of the originating station and the time intervals of train operation, time-domain standardization is performed, and the final result is... , , .

[0044] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A time-domain optimization method for urban rail transit based on the spatiotemporal distribution of passenger flow, characterized by: Includes the following steps: Step S1: Based on the urban rail network topology, using OD passenger flow as input, construct a passenger travel network, and use a multi-path search algorithm based on dynamic costs to extrapolate passenger spatiotemporal paths, reconstruct passenger travel paths, and statistically analyze the spatiotemporal distribution of passenger flow on the line. Step S2: Using the spatiotemporal distribution of passenger flow as input, perform time-domain optimization using ordered clustering and determine the number of time-domain optimizations. Step S3: Based on the principle of ordered sample clustering, construct a time-domain optimization framework for urban rail transit operation and perform bidirectional time-domain integration. Using the principle of ordered sample clustering, it is necessary to define the diameter of the cluster, as follows: Assumption yes An ordered sample, for dimensional vector; let one class Includes samples , recorded as The diameter of this class is defined as ; In the formula This represents the sum of squared deviations of the sample. The mean vector of this class ; Step 1: Using the characteristic day's OD passenger flow as input, and through the passenger flow spatiotemporal path extrapolation framework, statistically analyze the time-sharing cross-sectional passenger flow of the selected route; Step 2: Initialize clustering parameters and samples; ① Read the topology data of the research line and the passenger flow data of the characteristic day's time section, and set the operation start time as... The operation will end at [time]. Minimum granularity of time domain partitioning Set constraints on the number of time-domain partitions. ; ②Time Sample In the formula Representing the time domain The passenger flow state vector of the inward-flow section, where To study the first line A cross-section in the time domain Similarly, the upstream time sample can be obtained from the cross-sectional passenger flow. ; Step 3: Select the optimal number of time domain partitions ; Introducing a classification loss function, its contents include: use Indicates will The ordered samples are divided into A certain way of classifying, denoted as: ; In the formula, For classification indexing, the classification loss function is defined as follows: ; exist , At a certain time, The smaller the value, the higher the similarity between categories, and the more reasonable the classification. The goal of the optimal segmentation scheme is to find the classification method that minimizes the classification loss function among all classification results. ;Right now: ; The formula indicates that The ordered samples are divided into The optimal segmentation scheme for a class is based on the previous an ordered sample Optimal segmentation: Given the optimal number of segments, the optimal segmentation scheme is obtained based on the above formula. ; Step 4: Standardize the unidirectional time domain partitioning results The specific steps are as follows: ① Based on the optimal time-domain partitioning results, find the time-domain maximum passenger flow cross-section, and calculate the time-domain characteristic passenger flow density based on the maximum cross-section passenger flow value according to the following formula; where Representing the time domain Characteristic passenger flow density, Representing the time domain Maximum cross-sectional passenger flow Representing the time domain The length, in units of : ; ② Using the maximum passenger flow section as the initial section, the time-domain division result is standardized based on the directed interval running time and station stopping time, where... Representing the time domain The start or end time, Representing the time domain The total stopping time of the train from its originating station to the reference section. Representing the time domain The total travel time of trains from their originating station to the reference section: ; ③ Based on the time-domain characteristic passenger flow density, determine whether the difference between adjacent time-domain intervals is less than the time-domain interval limit. If so, then the adjacent time-domain partitioning results are merged to complete the standardization; where Representing the time domain Train intervals, , These represent the minimum and maximum headway times for the selected route, respectively. This indicates the train's rated passenger capacity, expressed in people per car. Representing the time domain Expected train occupancy rate: ; Step 5: Integrate the bidirectional time domain partitioning results.

2. The urban rail transit time-domain optimization method based on passenger flow spatiotemporal distribution deduction according to claim 1, characterized in that: Step S1 further includes the following: A passenger travel path search network based on the urban rail network topology, including ordinary nodes, transfer nodes, direct arcs, and transfer arcs, is constructed. To consider the sensitivity of passenger travel to changes in travel time and number of transfers, a transfer arc cost calculation method based on path travel time and number of transfers is adopted to fully describe the impact of transfers on passenger travel. The following settings are defined: For the set of transfer arcs, The transfer penalty coefficient is the origin of the travel network. The destination is The path, passing through the arc Costs incurred The calculation method is shown in the following formula: ; Indicates the station index of the line; Indicates the originating station is The final station is The interval; Representing an interval The interval running time; Indicates from the station Arrive at the station The path cost.

3. The urban rail transit time-domain optimization method based on passenger flow spatiotemporal distribution deduction according to claim 2, characterized in that: K-short-circuit search of the travel network based on the passenger travel network is performed using the A* algorithm: The A* algorithm is an efficient path search algorithm for static networks. It uses the evaluation function as the key to the A* algorithm, expressed as follows: ,in This represents the distance from the starting point to the current node. Path cost, This represents the distance from the endpoint to the current node. Shortest path cost; set Used to represent the distance from the starting point to the current point. The path is used in the algorithm to avoid searching for a detour; the K-short-circuit search steps based on the passenger travel network are as follows: Step 1: Initialize the road network and establish a directed network for passenger travel; Step 2: Calculate the endpoint using the SPFA algorithm The shortest path to all other points in the network is obtained. ; set up For nodes The set of preceding nodes, Recorded as express Arc cost, From the starting point to the node The set of paths To verify the node queue, the specific steps are as follows: ① Initialization, let , , , , ; ② From the queue Extract the first element from the list and use it as the current node. ; ③ Loop through the current node Forward node set Take node ; ④ Calculate the arc cost according to the formula. Determine whether and If so, then update. and ; ⑤ Determine if the update node is in the queue. If it is not in the queue, add it to the queue; ⑥ Determine the set Has the loop ended? If so, remove from the queue. Remove node Otherwise, repeat steps ④-⑤; ⑦ Determine the set Is it If yes, the algorithm ends; otherwise, repeat steps ③-⑥. Step 3: Use the A* algorithm to search for K-short circuits between travel networks; set up For nodes The set of subsequent nodes, Recorded as express Arc cost, From the starting point to the node The set of paths Starting point to The The path cost of the shortest circuit. To verify the node queue, The specific steps are as follows: ① Initialization, let , , , ; ② From the queue Extract the first element from the list and use it as the current node. ; ③ Loop through the current node The set of subsequent nodes Take node ; ④ Determine if the node is If the arc cost is calculated according to the formula ,renew , , ; ⑤ Determine if the node is If so, then output the first... The sum of the costs of the arcs traversed along the short circuit path is denoted as: ,make Otherwise Add to queue middle; ⑥ Determine the set Has the loop ended? If so, remove from the queue. Remove node ,queue Sort in ascending order according to the valuation function; otherwise, repeat steps ④-⑤. ⑦ Determine the set Is it or If yes, the algorithm ends; otherwise, repeat steps ③-⑥. Step 4: Based on the correspondence between the travel network and the topology network, reconstruct... Between Short circuit.

4. The urban rail transit time-domain optimization method based on passenger flow spatiotemporal distribution deduction according to claim 1, characterized in that: Step S1 further includes the following: The Logit model is used to calculate the path selection probability, starting from... The destination is Passenger routes The proportion is expressed as: where This is a non-negative parameter, and its magnitude represents the passenger's familiarity with the route. The larger the value, the greater the probability that the passenger will choose that route. ； Starting point is The destination is Passengers choose the first Passenger flow along the route The calculation formula is as follows: ； Travel Network Arc Passenger flow The calculation formula is as follows, where Indicates passing through arc The set of paths; ; In actual travel, various factors can cause discrepancies between passenger flow distribution and actual travel patterns.

5. The urban rail transit time-domain optimization method based on passenger flow spatiotemporal distribution deduction according to claim 4, characterized in that: The passenger flow allocation algorithm is as follows: Step 1: Initialize the passenger travel network and start passenger flow allocation. Passenger flow allocation end time passenger flow allocation granularity The basic time domain set for passenger flow allocation is ; Step 2: Calculate the travel network Between Shortest path; Step 3: Calculate the selection ratio of each path; Step 4: Take For the current time domain, statistical analysis is conducted within each time domain. Passenger flow value ; Step 5: Take For the present Determine whether If so, then calculate. And assign them to the corresponding travel networks; Step 6: Determine if completed Loop, if so, then take Backward time domain If it is the current time domain, otherwise repeat step 5; Step 7: Calculate the arc flow of the travel network in each time domain, and obtain the passenger flow of the time section and the passenger flow of the time transfer based on the arc type.

6. A non-volatile storage medium, characterized in that, The non-volatile storage medium includes a stored program, wherein the program, when executed, controls the device where the non-volatile storage medium is located to perform the method described in any one of claims 1 to 5.

7. An electronic device, characterized in that, It includes a processor and a memory; the memory stores computer-readable instructions, and the processor is configured to execute the computer-readable instructions, wherein the computer-readable instructions, when executed, perform the method according to any one of claims 1 to 5.

Images