Public traffic scheduling optimization method driven by multi-modal time series data analysis

Abstract

The invention relates to the technical field of intelligent traffic management, and discloses a multi-modal time series data analysis-driven public traffic scheduling optimization method, which comprises the following steps of: acquiring multi-modal traffic data, performing time-space alignment preprocessing, and extracting a feature vector; fusing various feature vectors based on a cross-modal interactive learning method; and performing real-time prediction of the public traffic condition based on the multi-modal fusion feature vector to obtain a public traffic flow prediction value, and making a public traffic scheduling strategy based on the obtained public traffic flow prediction value. According to the method, multiple feature vectors are fused based on a cross-modal interactive learning method, the mutual relation between different types of data is effectively captured, the public transport scheduling strategy is formulated based on the particle swarm optimization algorithm, the global search and local search capabilities are effectively balanced, the performance of the PSO algorithm in the public transport vehicle scheduling problem can be effectively improved, and the public transport vehicle scheduling efficiency is improved. And a more efficient scheduling scheme is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent transportation management, and more specifically, to a public transportation scheduling optimization method driven by multi-modal time-series data analysis. Background Art

[0002] With the acceleration of the urbanization process, the public transportation system is facing unprecedented challenges. On the one hand, the growth and expansion of the urban population have led to a sharp increase in the demand for public transportation; on the other hand, problems such as traffic congestion and environmental pollution have become increasingly serious, making it an important task for urban development to improve the efficiency and service quality of the public transportation system. Traditional public transportation scheduling mainly relies on empirical rules and historical data, and this method seems powerless in dealing with dynamic traffic conditions and is difficult to meet the requirements of real-time and accuracy.

[0003] In this context, using advanced data analysis technology to optimize public transportation scheduling has become a research hotspot. In particular, multi-modal time-series data analysis, as an emerging technical means, can effectively integrate data from different sources. However, how to effectively fuse these different types of data and extract useful information from them to support decision-making is a complex problem.

[0004] Therefore, how to provide a public transportation scheduling optimization method driven by multi-modal time-series data analysis is an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0005] In view of this, the present invention provides a public transportation scheduling optimization method driven by multi-modal time-series data analysis, which can not only improve the response speed and accuracy of the public transportation system, but also enhance its ability to adapt to dynamic environments, thereby providing more efficient and convenient services for passengers.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] A public transportation scheduling optimization method driven by multi-modal time-series data analysis, comprising:

[0008] Obtaining multi-modal traffic data and performing spatio-temporal alignment preprocessing to extract feature vectors;

[0009] Fusing multiple feature vectors based on a cross-modal interaction learning method;

[0010] Performing real-time prediction of the public transportation situation based on the multi-modal fusion feature vectors to obtain the public transportation flow prediction value;

[0011] Formulating a public transportation scheduling strategy based on the public transportation flow prediction value.

[0012] Preferably, the multi-modal traffic data includes in-vehicle satellite positioning data, traffic flow data, and real-time data feedback from social software. Among them, the structured data is traffic flow data, the semi-structured data includes in-vehicle satellite positioning data, and the unstructured data includes text data feedback from social software.

[0013] Preferably, the extraction of feature vectors includes:

[0014] Processing the structured data based on an autoencoder;

[0015] Constructing an urban traffic road network map, and extracting spatial features from the semi-structured data based on a graph attention network. The adjacency matrix of the graph attention network is constructed according to the road topology relationship;

[0016] Performing time series analysis and key information extraction on the unstructured data based on a bidirectional long short-term memory network.

[0017] Preferably, fusing multiple feature vectors based on a cross-modal interaction learning method, including:

[0018] (1) Constructing a multi-modal spatio-temporal association graph based on the urban traffic road network map. The nodes represent different modal feature vectors, and the edge weights are calculated through a spatio-temporal attention mechanism. The process of calculating the edge weights is as follows:

[0019] Calculating the modal attention weights:

[0020]

[0021] Among them, respectively represent the query transformation matrix and the key transformation matrix. K represents the number of attention heads, D is the transformed feature dimension, and respectively represent the feature vectors of node i and node j, is the modal attention weight;

[0022] Calculating the attention weight calculated based on the spatial distance:

[0023]

[0024] Among them, represents the Euclidean distance of the grid center point, is the time decay coefficient, represents the attention weight calculated based on the spatial distance;

[0025] Calculating the attention weight calculated based on the time difference:

[0026]

[0027] Among them, represents the time decay coefficient, represents the time interval between two nodes, represents the attention weight calculated based on the time difference;

[0028] Normalized attention weight:

[0029]

[0030] where, represents the activation function, represents the final attention weight from node i to node j;

[0031] (2) Aggregate multi-modal features through the graph attention network, including:

[0032] Use K independent attention heads to calculate in parallel:

[0033]

[0034] where, , represents the value transformation matrix of the k-th attention head, represents the neighbor nodes of node i, represents the normalized attention weight of the k-th attention head, represents the feature of the k-th attention head;

[0035] After calculating in parallel for multiple attention heads, perform multi-head feature splicing to output the intermediate fusion feature;

[0036] Perform residual connection and normalization on the intermediate fusion feature;

[0037] (3) After aggregating multi-modal features through the graph attention network, output node features through multiple layers of spatio-temporal graph convolution;

[0038] The final node features are mapped to a multi-modal fusion feature vector through a fully connected layer.

[0039] Preferably, perform real-time prediction of public transportation conditions based on the multi-modal fusion feature vector, including:

[0040] Use the spatio-temporal graph convolution network to jointly extract spatio-temporal features of the fusion feature;

[0041] Adopt a multi-scale time window mechanism to capture time series dependency relationships at different granularities;

[0042] Dynamically adjust the feature weights through the spatio-temporal gated unit to output the public transportation flow prediction value.

[0043] Preferably, formulate a public transportation scheduling strategy based on the particle swarm optimization algorithm, including:

[0044] Take the multi-modal fusion feature vector as the initial position and define the adaptability function;

[0045] Calculate the particle fitness based on the predicted value of public transportation flow;

[0046] Update the position and velocity of each particle according to the particle fitness, and consider the spatio-temporal constraint conditions during the update process;

[0047] Terminate when the update process reaches the maximum number of iterations or the fitness change is less than the set threshold to obtain the global optimal solution, which is the public transportation scheduling strategy.

[0048] Preferably, the update formulas for the position and velocity of each particle are as follows:

[0049]

[0050] Among them, is the updated velocity, is the velocity of particle n in the m-th dimension, is the position of particle n in the m-th dimension, is the contraction factor, and are the maximum and minimum values of the inertia weight respectively, t is the current iteration number, is the maximum number of iterations, are the initial and final learning factors respectively values, are the initial and final learning factors respectively values, is the historical optimal position of particle n in the m-th dimension, is the optimal position of the entire particle swarm in the m-th dimension, and are random numbers between [0, 1];

[0051]

[0052] Among them, is the updated position.

[0053] Preferably, the public transportation scheduling strategy includes:

[0054] Departure time of the vehicle: The time point when each vehicle departs from the starting station;

[0055] Number of vehicles: The number of vehicles to be put into a certain line;

[0056] Running frequency of the vehicle: The interval time between vehicles;

[0057] Running route of the vehicle: Whether the vehicle needs to detour certain sections or adjust the running route;

[0058] Station dwell time: The time a vehicle stays at each station.

[0059] A computer device, characterized in that it includes: a memory and a processor, wherein a computer program that can run on the processor is stored in the memory, and when the processor executes the computer program, a public transportation scheduling optimization method driven by multi-modal time series data analysis is implemented.

[0060] A computer-readable storage medium, characterized in that a computer program is stored on the storage medium, and when the computer program is executed by a processor, a public transportation scheduling optimization method driven by multi-modal time series data analysis is implemented.

[0061] As can be seen from the above technical solutions, compared with the prior art, the present invention discloses a public transportation scheduling optimization method driven by multi-modal time series data analysis, which fuses multiple feature vectors based on a cross-modal interaction learning method, effectively captures the mutual relationship between different types of data, and formulates a public transportation scheduling strategy based on a particle swarm optimization algorithm, effectively balancing the global search and local search capabilities, realizing a more efficient scheduling scheme, and at the same time forming a technical context of "spatiotemporal alignment - spatiotemporal feature extraction - spatiotemporal attention fusion - spatiotemporal joint prediction - constrained optimization scheduling" as a whole. Therefore, the present invention not only realizes the all-round perception and understanding of the urban public transportation system, but also effectively improves the scientificity and rationality of scheduling decisions, and finally achieves the purpose of improving transportation efficiency and service quality. Brief Description of the Drawings

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings according to the provided drawings without creative efforts.

[0063] Figure 1 It is a flowchart of a public transportation scheduling optimization method driven by multi-modal time series data analysis provided by the present invention.

[0064] Figure 2 It is a flowchart of fusing multiple feature vectors based on a cross-modal interaction learning method provided by the present invention.

[0065] Figure 3 It is a flowchart of formulating a public transportation scheduling strategy based on a particle swarm optimization algorithm provided by the present invention. Detailed Embodiments

[0066] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0067] An embodiment of the present invention discloses a public transportation scheduling optimization method driven by multi-modal time-series data analysis, as Figure 1 shown, including:

[0068] Obtain multi-modal traffic data and perform spatio-temporal alignment preprocessing, and extract feature vectors. The modal traffic data includes in-vehicle satellite positioning data, traffic flow data, and real-time data feedback by social software. In addition to the above data, other modal traffic data can also be obtained according to actual needs. Among them, the structured data is traffic flow data, including road segment ID, statistical time window (such as every 5 minutes), number of vehicles, average speed, congestion level, etc. The semi-structured data includes in-vehicle satellite positioning data, and the satellite positioning data contains information such as vehicle ID, longitude and latitude coordinates, timestamp, instantaneous speed, direction, etc. The unstructured data includes text data feedback by social software, including but not limited to text content such as traffic accidents or emergencies occurring on the road, large event previews, etc. shared through social software, release time, and geographical location tags; The spatio-temporal alignment preprocessing includes a time alignment module and a space alignment module. Among them, the time alignment module uses the dynamic time warping algorithm to align the timestamps of multi-modal data, and the space alignment module unifies the spatial coordinate system through geographic grid coding.

[0069] Time-align asynchronous data:

[0070] Take the traffic flow data as the reference time axis (for example, a fixed 5-minute interval);

[0071] Generate a coordinate sequence aligned with the reference time axis for the in-vehicle satellite positioning data by linear interpolation;

[0072] Aggregate the social media text by time window, distribute the text to the nearest reference time window according to the release time, and extract event features.

[0073] Spatial alignment:

[0074] Define the urban geographic grid (such as a 500m×500m grid):

[0075] In the in-vehicle satellite positioning data, convert the longitude and latitude of each trajectory point into a grid ID, and count the number of vehicles, average speed, etc. in each grid.

[0076] In traffic flow data, map the road segment ID to the set of covered grids (for example, road segment A spans grids G0235 and G0236), and distribute the flow values according to the length ratio.

[0077] For social text data, perform location entity recognition, that is, use the Geographical Named Entity Recognition (GeoNER) model to extract locations (such as "Intersection A") in the text and associate them with grid IDs. For texts without clear location descriptions, map them according to the user's IP or the published location tag.

[0078] The present invention solves the problems of time asynchrony and spatial heterogeneity through dynamic time warping and geographical grid coding.

[0079] Fuse multiple feature vectors based on the cross-modal interaction learning method;

[0080] Based on the multi-modal fusion feature vectors, perform real-time prediction of the public transportation status to obtain the public transportation flow prediction value;

[0081] Formulate a public transportation scheduling strategy based on the public transportation flow prediction value.

[0082] In this embodiment, extract feature vectors, including:

[0083] Extract feature vectors from the preprocessed structured data based on an autoencoder. Specifically, the autoencoder learns a low-dimensional representation form through encoding and decoding operations on the structured data. This feature vector has the characteristic of noise reduction and can better reflect the true distribution of the data;

[0084] Construct an urban traffic road network map, and perform spatial feature extraction on semi-structured data based on a graph attention network to obtain feature vectors. The adjacency matrix of the graph attention network is constructed according to the road topology relationship; the specific process of constructing the urban traffic road network map is as follows:

[0085] Obtain urban road network data, usually from open map services (such as OpenStreetMap) or GIS data of traffic management departments.

[0086] Take the key points of intersections or road segments as the nodes of the graph, and each node contains longitude and latitude coordinates. For example, set the starting point / ending point of each intersection or road segment as a node.

[0087] Define edges according to the actual connection situation of the roads. For example, if two roads are connected at an intersection, establish an edge between the corresponding nodes.

[0088] Add attributes to the nodes and edges, such as node coordinates, road types (main roads, secondary roads), number of lanes, speed limits, etc.

[0089] The road segments and intersections of the urban traffic road network map will be mapped to the corresponding geographical grids.

[0090] Time series analysis and key information extraction are performed on the preprocessed unstructured data based on a bidirectional long short-term memory network, and the specific processing process is also implemented through existing technologies.

[0091] In this embodiment, a cross-modal interaction learning method is used to fuse multiple feature vectors. In the cross-modal interaction learning method, fusing multiple feature vectors to generate a richer and more comprehensive multi-modal representation is a core step. As Figure 2 shown, it specifically includes:

[0092] Each spatio-temporal node corresponds to a geographical grid i and a time window t. The node represents different modal feature vectors, and the edge weights are calculated through a spatio-temporal attention mechanism. The calculation formula is:

[0093]

[0094] Among them, respectively represent the query transformation matrix and the key transformation matrix. K represents the number of attention heads, and D is the dimension of the transformed features. and respectively represent the feature vectors of node i and node j. is the modal attention weight;

[0095]

[0096] Among them, represents the Euclidean distance of the grid center point. is the time decay coefficient. represents the attention weight calculated based on the spatial distance;

[0097]

[0098] Among them, represents the time decay coefficient. represents the time interval between two nodes. represents the attention weight calculated based on the time difference;

[0099]

[0100] Among them, represents the activation function. represents the final attention weight from node i to node j;

[0101] Aggregate multi-modal features through a graph attention network, specifically including:

[0102] Use K independent attention heads to calculate in parallel:

[0103]

[0104] Among them, , represents the value transformation matrix of the k-th attention head, represents the neighbor nodes of node i, represents the normalized attention weight of the k-th attention head, represents the feature of the k-th attention head;

[0105] After parallel computing for multiple attention heads, multi-head feature splicing is performed to output intermediate fusion features;

[0106] Residual connection and normalization are performed on the intermediate fusion features;

[0107] After aggregating multi-modal features through the Graph Attention Network (GAT), node features are output through multi-layer spatio-temporal graph convolution;

[0108] The final node features are mapped to a multi-modal fusion feature vector through a fully connected layer.

[0109] Cross-modal interaction learning is an advanced data processing method that enhances the expressiveness of the model by integrating information from different modalities (such as structured data, semi-structured data, and unstructured data). And the triple attention mechanism of modality + space + time in the present invention deeply fuses multi-modal data in a unified spatio-temporal graph structure, while inheriting the physical constraints of the road network graph, providing a feature vector with strong representation ability for subsequent prediction and optimization, and enhancing the generalization ability and robustness of the model in the time and space dimensions.

[0110] In this embodiment, real-time prediction of public transportation conditions is performed based on the multi-modal fusion feature vector, including:

[0111] Using the Spatio-Temporal Graph Convolutional Network (STGCN) to jointly extract spatio-temporal features of the fusion features;

[0112] Adopting a multi-scale time window mechanism to capture temporal dependence relationships at different granularities;

[0113] Dynamically adjusting feature weights through a spatio-temporal gated unit to output the predicted value of public transportation flow.

[0114] Based on graph attention aggregation, the present invention further uses multi-layer spatio-temporal graph convolution (ST-GCN) to extract deep spatio-temporal joint features, significantly improving the accuracy and timeliness of traffic flow prediction. At the same time, combined with a spatio-temporal gated unit to adaptively adjust the weights of each branch, the adaptability of the model to complex traffic patterns is improved.

[0115] In this embodiment, a public transportation scheduling strategy is formulated based on the particle swarm optimization algorithm, as Figure 3 shown, including:

[0116] Take the multi-modal fusion feature vector as the initial position, and randomly initialize the velocity of each particle. The position of each particle (representing different scheduling schemes, such as the number of vehicles at different time points, route selection, etc.);

[0117] Design a comprehensive fitness function that combines factors such as vehicle cost, passenger waiting time, and load balance, which can be:

[0118]

[0119] Among them, is the weight coefficient, which can be adjusted according to the actual situation; , , respectively represent vehicle cost, passenger waiting time, and route load balance;

[0120] Calculate the particle fitness based on the predicted value of public transportation flow;

[0121]

[0122] Among them, is the updated velocity, is the velocity of particle n in dimension m, is the position of particle n in dimension m, is the contraction factor, and are the maximum and minimum values of the inertia weight respectively, t is the current iteration number, is the maximum iteration number, are the initial and final learning factors respectively values of, are the initial and final learning factors respectively values of, is the historical optimal position of particle n in the m-th dimension, is the optimal position of the entire particle swarm in the m-th dimension, and are random numbers between [0, 1];

[0123]

[0124] Among them, is the updated position.

[0125] Consider spatio-temporal constraint conditions during the update process, such as:

[0126] Maximum empty running distance: Limit the maximum distance that a vehicle can travel without passengers.

[0127] Minimum headway: Ensure there is sufficient time interval between consecutive departures at the same station.

[0128] These constraints can be achieved by adding penalty terms to the fitness function. For example, if a particle violates the minimum headway constraint, a large penalty value can be added to its fitness.

[0129] The iterative update process terminates when the maximum number of iterations is reached or the fitness change is less than the set threshold, obtaining the global optimal solution, which is the public transportation scheduling strategy.

[0130] The inertia weight of the present invention controls the degree of maintaining the original speed in the speed update of the particle. A larger inertia weight is helpful for global search, while a smaller inertia weight is helpful for local fine search. The present invention adjusts the inertia weight in a linearly decreasing manner, effectively balancing the global search and local search capabilities. At the same time, in order to prevent the search from being unstable due to excessive particle speed, the present invention introduces a contraction factor to limit the speed, and combines with an improved particle swarm optimization algorithm (PSO) to minimize the operation cost, passenger waiting time, and load balance, while meeting a series of constraints in actual operations (such as the maximum deadhead distance, minimum headway, etc.), to formulate the optimal bus scheduling plan, which can effectively improve the performance of the PSO algorithm in the public transportation vehicle scheduling problem.

[0131] Mapping the optimal solution to the actual scheduling operation may include the following:

[0132] Departure time adjustment: Adjust the departure schedules of each line according to the departure time information in the optimal solution.

[0133] Vehicle number adjustment: Increase or decrease the number of vehicles put into operation according to the vehicle number information in the optimal solution.

[0134] Running frequency adjustment: Adjust the headway between vehicles according to the running frequency information in the optimal solution.

[0135] Running route adjustment: Adjust the running routes of vehicles according to the route information in the optimal solution, such as bypassing congested sections or adding temporary routes.

[0136] Station stop time adjustment: Optimize the stop time of vehicles at each station according to the station stop time information in the optimal solution.

[0137] This embodiment is not limited thereto, and the specific scheduling content can be dynamically adjusted according to actual needs.

[0138] This embodiment provides a computer device, including: a memory and a processor. A computer program that can run on the processor is stored in the memory. When the processor executes the computer program, a public transportation scheduling optimization method driven by multi-modal time series data analysis is implemented.

[0139] This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, a public transportation scheduling optimization method driven by multi-modal time series data analysis is implemented.

[0140] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps including the above method embodiments; and the foregoing storage medium includes: various media that can store program codes, such as removable storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0141] The various embodiments in this specification are described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0142] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A public transportation scheduling optimization method driven by multi-modal time series data analysis, characterized in that include: Acquire multimodal traffic data and perform spatiotemporal alignment preprocessing to extract feature vectors; Fusion of multiple feature vectors based on cross-modal interactive learning method; Based on the multimodal fusion feature vector, the public transportation status is predicted in real time to obtain the public transportation flow prediction value; Formulate public transportation scheduling strategies based on public transportation flow forecast values.

2. The multimodal time-series data analysis-driven public transportation scheduling optimization method according to claim 1, wherein Multimodal traffic data includes vehicle-mounted satellite positioning data, traffic flow data, and real-time data from social software feedback. The structured data is traffic flow data, the semi-structured data includes vehicle-mounted satellite positioning data, and the unstructured data includes text data from social software feedback.

3. A public transportation scheduling optimization method driven by multimodal time-series data analysis according to claim 2, characterized in that Extracting feature vectors includes: Processing structured data based on autoencoders; Construct an urban traffic road network map and extract spatial features from semi-structured data based on a graph attention network. The adjacency matrix of the graph attention network is constructed based on the road topology relationship. Time series analysis and key information extraction of unstructured data based on bidirectional long short-term memory network.

4. A public transportation scheduling optimization method driven by multimodal time series data analysis according to claim 3, characterized in that, Based on the cross-modal interactive learning method, multiple feature vectors are integrated, including: (1) A multimodal spatiotemporal association graph is constructed based on the urban traffic road network graph. The nodes represent the feature vectors of different modalities. The edge weights are calculated through the spatiotemporal attention mechanism. The edge weight calculation process is: Calculate the modality attention weight: ; Among them, respectively represent the query transformation matrix and the key transformation matrix, K represents the number of attention heads, and D is the dimension of the transformed features. and respectively represent the feature vectors of nodes i and j. is the modal attention weight. Calculate the attention weight based on the spatial distance calculation: ; Among them, represents the Euclidean distance of the grid center point, is the time decay coefficient, represents the attention weight calculated based on the spatial distance; Calculate the attention weight based on the time difference: ; Among them, represents the time decay coefficient, represents the time interval between two nodes, represents the attention weight calculated based on the time difference; Normalized attention weights: ; Among them, represents the activation function, represents the final attention weight from node i to node j ; (2) Aggregate multimodal features through graph attention networks, including: Use K independent attention heads to parallelize the computation: ; Among them, represents the value transformation matrix of the k -th attention head, represents the neighbor nodes of node i , represents the normalized attention weight of the k -th attention head, represents the feature of the k -th attention head; After parallel calculation of multiple attention heads, multiple features are concatenated to output intermediate fusion features. Perform residual connection and normalization on the intermediate fusion features; (3) After the multimodal features are aggregated by the graph attention network, the node features are output through multiple layers of spatiotemporal graph convolution. The node features are mapped into multimodal fusion feature vectors through the fully connected layer.

5. A public transportation scheduling optimization method driven by multimodal time-series data analysis according to claim 1, characterized in that Real-time prediction of public transportation conditions based on multimodal fusion feature vectors, including: Use the spatiotemporal graph convolutional network to jointly extract spatiotemporal features from the fused features; A multi-scale time window mechanism is used to capture temporal dependencies of different granularities; The feature weights are dynamically adjusted through the spatiotemporal gating unit to output the public transportation flow prediction value.

6. The multimodal time-series data analysis-driven public transport scheduling optimization method according to claim 1, wherein Formulate public transportation scheduling strategies based on particle swarm optimization algorithm, including: The multimodal fusion feature vector is used as the initial position and the adaptability function is defined; Calculate particle fitness based on public transportation flow prediction values; Update the position and velocity of each particle according to its fitness, taking into account the space-time constraints during the update process; The update process is terminated when the maximum number of iterations is reached or the fitness change is less than the set threshold, and the global optimal solution is obtained. The global optimal solution is the public transportation scheduling strategy.

7. A public transport scheduling optimization method driven by multimodal time series data analysis according to claim 6, characterized in that The update formula for the position and velocity of each particle is: ; Among them, is the updated speed, is the particle n in the m dimensional speed, is the particle n in the m dimensional position, is the contraction factor, and are the maximum and minimum values of the inertia weight respectively, t is the current iteration number, is the maximum iteration number, are the initial and final learning factors respectively values, are the initial and final learning factors respectively values, is the particle n in the m dimensional historical optimal position, is the optimal position of the entire particle swarm in the m dimensional, and are random numbers between [0, 1]; ; Among them, is the updated position.

8. A public transportation scheduling optimization method driven by multimodal time series data analysis according to claim 1, characterized in that Public transport scheduling strategies include: Vehicle departure time: the time when each vehicle departs from the departure station; Number of vehicles: the number of vehicles required on a certain route; The operating frequency of vehicles: the interval time between vehicles; Vehicle operation path: whether the vehicle needs to detour certain sections of the road or adjust the operation route; Stop time at each stop: the time the vehicle stays at each stop.

9. A computer device, characterized in that, Comprising: A memory and a processor, wherein a computer program that can run on the processor is stored in the memory, and when the processor executes the computer program, the method according to any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium, characterized in that, A computer program is stored on the storage medium, and when the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images