Resource scheduling method and device of subway system, equipment and medium

Abstract

The invention relates to the technical field of rail transit, and provides a resource scheduling method and device for a subway system, equipment and a medium. The method comprises the steps of obtaining first multi-modal data corresponding to a subway system, and performing data fusion based on the first multi-modal data to generate an associated database; determining the device priority of each device in the subway system according to the association database; constructing a three-dimensional resource scheduling model corresponding to the subway system based on the space dimension, the time dimension and the cost dimension; performing collaborative optimization on the three-dimensional resource scheduling model based on the device priority of each device to obtain an optimized three-dimensional resource scheduling model; and generating a resource scheduling scheme according to the optimized three-dimensional resource scheduling model, and performing resource scheduling on the subway system according to the resource scheduling scheme. Based on the scheme of the invention, the resource scheduling accuracy of the subway system can be improved.

Description

Technical Field

[0001] This invention relates to the field of rail transit technology, and in particular to a resource scheduling method, apparatus, equipment and medium for a subway system. Background Technology

[0002] With the continuous expansion of urban subway networks, passenger flow is exhibiting a networked distribution pattern. The resource scheduling schemes of related subway systems mainly rely on static models based on fixed timetables and human experience. These schemes are difficult to adjust accurately in the face of tidal changes in passenger flow and equipment failures, easily leading to reduced operational efficiency, increased maintenance costs, and cross-line resource conflicts. Therefore, improving the accuracy of resource scheduling in subway systems has become an urgent technical problem to be solved. Summary of the Invention

[0003] This application provides a resource scheduling method, apparatus, equipment, and medium for a subway system, to address the technical problem of improving the accuracy of resource scheduling in a subway system.

[0004] In a first aspect, embodiments of this application provide a resource scheduling method for a subway system, including: Obtain the first multimodal data corresponding to the subway system, and perform data fusion based on the first multimodal data to generate a related database; Determine the equipment priority of each device in the subway system based on the associated database; Based on spatial, temporal, and cost dimensions, a three-dimensional resource scheduling model for the subway system is constructed. Based on the device priority of each device, the 3D resource scheduling model is collaboratively optimized to obtain the optimized 3D resource scheduling model. A resource scheduling scheme is generated based on the optimized 3D resource scheduling model, and the subway system is then scheduled according to the resource scheduling scheme.

[0005] In conjunction with the first aspect, in some possible implementations, the first multimodal data corresponding to the subway system is obtained, including: Acquire the equipment status data corresponding to the subway system. The equipment status data is used to characterize the real-time operating conditions and fault trends of the equipment in the subway system. Obtain passenger flow data corresponding to the subway system. Passenger flow data is used to characterize the spatiotemporal distribution and congestion situation of passenger groups in the subway system. Obtain environmental data corresponding to the subway system. The environmental data is used to characterize the parameter changes and influencing factors of environmental conditions in the subway system. Based on the equipment status data, passenger flow data, and environmental data corresponding to the subway system, the first multimodal data corresponding to the subway system is determined.

[0006] Combining the first aspect and the above implementation methods, in some possible implementation methods, data fusion is performed based on the first multimodal data to generate a related database, including: The station-level edge computing nodes of the subway system are invoked to preprocess the first multimodal data to obtain the second multimodal data; Data fusion is performed based on the second multimodal data to generate a related database.

[0007] Combining the first aspect and the above implementation methods, in some possible implementation methods, data fusion is performed based on the second multimodal data to generate a related database, including: Data cleaning is performed on the second multimodal data to obtain cleaned second multimodal data; Feature data is obtained by extracting features from the cleaned second multimodal data; The feature data is analyzed based on a pre-defined long short-term memory network to obtain prediction results; the prediction results are used to characterize the impact of equipment failure on passenger flow prediction. Based on the prediction results and characteristic data, establish the mapping relationship between equipment and passenger flow in the subway system; Based on the mapping relationship, a related database is generated.

[0008] Combining the first aspect and the above implementation methods, in some possible implementation methods, a three-dimensional resource scheduling model corresponding to the subway system is constructed based on the spatial, temporal, and cost dimensions, including: Based on the geographical distribution of resources and network topology of the subway system, a first model corresponding to the spatial dimension is constructed. Based on the resource scheduling time, task execution time, and resource availability time of the subway system, a second model corresponding to the time dimension is constructed. Based on the resource procurement cost, resource transportation cost, resource storage cost, and resource operation cost of the subway system, a third cost-dimensional model is constructed. Based on the first model, the second model, and the third model, a three-dimensional resource scheduling model corresponding to the subway system is constructed.

[0009] Combining the first aspect and the above implementation methods, in some possible implementation methods, the 3D resource scheduling model is collaboratively optimized based on the device priority of each device to obtain an optimized 3D resource scheduling model, including: Based on the device priority of each device, the optimization weight coefficients corresponding to the spatial dimension, time dimension and cost dimension in the three-dimensional resource scheduling model are dynamically adjusted to obtain the adjusted optimization weight coefficients; Based on the adjusted optimization weight coefficients, the three-dimensional resource scheduling model is solved through collaborative optimization to generate an optimized three-dimensional resource scheduling model.

[0010] Combining the first aspect and the above implementation methods, in some possible implementation methods, a resource scheduling scheme is generated based on the optimized three-dimensional resource scheduling model, including: The optimized three-dimensional resource scheduling model is solved to generate an initial resource scheduling instruction set; wherein, the initial resource scheduling instruction set includes at least one of train operation scheduling instructions, equipment maintenance scheduling instructions, and resource configuration scheduling instructions; The station-level edge computing nodes of the subway system are invoked to globally coordinate the initial resource scheduling instruction set based on a preset distributed consensus algorithm, resulting in a coordinated resource scheduling instruction set. Based on the coordinated resource scheduling instruction set, a resource scheduling scheme is generated.

[0011] Secondly, embodiments of this application provide a resource scheduling device for a subway system, comprising: The acquisition module is used to acquire the first multimodal data corresponding to the subway system, and perform data fusion based on the first multimodal data to generate a related database; The determination module is used to determine the equipment priority of each device in the subway system based on the associated database. The building module is used to construct a three-dimensional resource scheduling model for the subway system based on spatial, temporal, and cost dimensions. The optimization module is used to collaboratively optimize the 3D resource scheduling model based on the device priority of each device, so as to obtain the optimized 3D resource scheduling model. The scheduling module is used to generate a resource scheduling plan based on the optimized three-dimensional resource scheduling model, and to perform resource scheduling on the subway system according to the resource scheduling plan.

[0012] Thirdly, embodiments of this application provide an electronic device, including a processor and a memory storing a computer program, wherein the processor executes the program to implement the steps of the resource scheduling method for a subway system of the first aspect.

[0013] Fourthly, embodiments of this application provide a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the resource scheduling method for the subway system of the first aspect.

[0014] The resource scheduling method, apparatus, equipment, and medium for a subway system provided in this application first acquire the first multimodal data corresponding to the subway system, and then perform data fusion based on the first multimodal data to generate a relational database. Next, the equipment priority of each device in the subway system is determined based on the relational database. Then, a three-dimensional resource scheduling model corresponding to the subway system is constructed based on spatial, temporal, and cost dimensions. Then, the three-dimensional resource scheduling model is collaboratively optimized based on the equipment priorities of each device to obtain an optimized three-dimensional resource scheduling model. Finally, a resource scheduling scheme is generated based on the optimized three-dimensional resource scheduling model, and resource scheduling of the subway system is performed according to this scheme. Through the above process, the relational database generated based on the first multimodal data provides a comprehensive data foundation for determining the equipment priorities of each device; the three-dimensional resource scheduling model, including spatial, temporal, and cost dimensions, establishes multi-dimensional optimization objectives for collaborative optimization; furthermore, the collaborative optimization of the three-dimensional resource scheduling model based on equipment priorities enables the generated resource scheduling scheme to comprehensively consider multi-dimensional factors and key equipment requirements, thereby improving the accuracy of resource scheduling in the subway system. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating the resource scheduling method for a subway system provided in an embodiment of this application; Figure 2 This is a schematic diagram of the structure of the resource scheduling device for the subway system provided in the embodiments of this application; Figure 3 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0018] Currently, urban subway networks are expanding rapidly, and passenger flow is exhibiting a networked distribution pattern. Resource scheduling schemes for these subway systems primarily rely on static models based on fixed timetables and human experience.

[0019] Specifically, in some related technologies, the scheduling scheme is mainly based on a preset train timetable, which is generated based on historical passenger flow data and includes parameters such as fixed stop times and departure intervals. In this mode, the subway system's key resources, such as vehicles, signaling systems, and power, are usually allocated according to preset proportions. For example, a fixed number of reserve trains are deployed during peak hours, while capacity is reduced during off-peak hours.

[0020] In other related technologies, resource scheduling employs centralized control by a dispatch center, tracking trains via radio communication and block systems. Due to a lack of real-time data analysis capabilities, when the subway system experiences unexpected events such as equipment failures or abnormal passenger flow, the dispatch response primarily relies on the dispatcher's manual judgment and experience-based adjustments.

[0021] It is evident that the scheduling modes of related technologies struggle to accurately adjust to scenarios such as tidal changes in passenger flow and equipment malfunctions, easily leading to reduced operational efficiency, increased maintenance costs, and cross-line resource conflicts. Therefore, improving the accuracy of resource scheduling in subway systems has become an urgent technical problem to be solved.

[0022] To address the aforementioned issues, the main solution provided in this application includes: firstly, acquiring first multimodal data corresponding to the subway system, and then performing data fusion based on the first multimodal data to generate a relational database; subsequently, determining the equipment priority of each device in the subway system based on the relational database; next, constructing a three-dimensional resource scheduling model corresponding to the subway system based on spatial, temporal, and cost dimensions; then, collaboratively optimizing the three-dimensional resource scheduling model based on the equipment priority of each device to obtain an optimized three-dimensional resource scheduling model; finally, generating a resource scheduling scheme based on the optimized three-dimensional resource scheduling model, and performing resource scheduling on the subway system according to the resource scheduling scheme. Through the above process, the relational database generated based on the first multimodal data provides a comprehensive data foundation for determining the equipment priority of each device; the construction of a three-dimensional resource scheduling model including spatial, temporal, and cost dimensions establishes multi-dimensional optimization objectives for collaborative optimization; furthermore, the collaborative optimization of the three-dimensional resource scheduling model based on equipment priority enables the generated resource scheduling scheme to comprehensively consider multi-dimensional factors and key equipment requirements, thereby improving the accuracy of resource scheduling in the subway system.

[0023] The resource scheduling method for a subway system provided in the embodiments of this application will be described in detail below.

[0024] Please see Figure 1 , Figure 1 This is a flowchart illustrating a resource scheduling method for a subway system provided in an embodiment of this application. Figure 1 As shown, the method in this application embodiment may include the following steps S101-S105.

[0025] S101, Obtain the first multimodal data corresponding to the subway system, and perform data fusion based on the first multimodal data to generate a related database.

[0026] Specifically, the first step is to acquire the first multimodal data corresponding to the subway system, and then perform data fusion based on this first multimodal data to generate a relational database. Here, the subway system refers to the urban rail transit network; the first multimodal data corresponding to the subway system includes equipment status data, passenger flow data, and environmental data; data fusion refers to integrating and processing the equipment status data, passenger flow data, and environmental data; and the relational database refers to a database that stores the mapping relationship between equipment and passenger flow influences.

[0027] Regarding this step, in some possible implementations, relevant data processing techniques can be used to process the first multimodal data, and the processing results can be fused to generate a relational database. In some possible implementations, the first multimodal data can be obtained through a preset data interface, and relevant fusion algorithms can be applied to integrate the first multimodal data, thereby establishing a relational database.

[0028] S102, determine the equipment priority of each device in the subway system based on the associated database.

[0029] Specifically, in order to quantify the importance of equipment in resource scheduling and optimize allocation efficiency, it is necessary to determine the equipment priority of each piece of equipment in the subway system based on the associated database. Here, equipment in the subway system refers to key facilities such as trains, power supply systems, and ventilation and air conditioning; the equipment priority refers to a priority value calculated based on historical failure frequency, passenger flow impact coefficient, and maintenance cost weight.

[0030] Regarding this step, some possible implementations involve using historical operational data from a relational database and performing relevant mathematical operations to calculate equipment priorities for each device in the subway system. Other possible implementations involve dynamically adjusting the equipment priorities of each device in the subway system using relevant predictive models based on real-time status data from the relational database.

[0031] S103 constructs a three-dimensional resource scheduling model for the subway system based on spatial, temporal, and cost dimensions.

[0032] Specifically, in order to establish a multi-dimensional optimization framework to balance geographical, temporal, and economic factors in resource allocation, a three-dimensional resource scheduling model corresponding to the subway system needs to be constructed based on spatial, temporal, and cost dimensions. The spatial dimension refers to the geographical distribution of resources and the network topology; the temporal dimension refers to resource scheduling time, task execution time, and resource availability time; the cost dimension refers to resource procurement costs, transportation costs, storage costs, and operating costs; and the three-dimensional resource scheduling model is a comprehensive optimization model integrating the spatial, temporal, and cost dimensions.

[0033] Regarding this step, some possible implementations involve establishing corresponding sub-models for the spatial, temporal, and cost dimensions, and then integrating these sub-models to construct a three-dimensional resource scheduling model. Alternatively, relevant modeling methods can be applied, using the spatial, temporal, and cost dimensions as input parameters to directly construct the three-dimensional resource scheduling model.

[0034] S104. Based on the device priority of each device, the three-dimensional resource scheduling model is collaboratively optimized to obtain the optimized three-dimensional resource scheduling model.

[0035] Specifically, in order to dynamically adjust the model weights to improve the overall system performance, it is necessary to perform collaborative optimization of the 3D resource scheduling model based on the device priorities of each device, resulting in an optimized 3D resource scheduling model. Collaborative optimization refers to adjusting the optimization weight coefficients of the spatial, temporal, and cost dimensions in the 3D resource scheduling model according to device priorities; the optimized 3D resource scheduling model refers to the scheduling model after weight adjustment and solution.

[0036] Regarding this step, in some possible implementations, the weight parameters related to the spatial, temporal, and cost dimensions in the 3D resource scheduling model can be adjusted based on the device priorities of each device. The adjusted model can then be solved using relevant optimization algorithms to obtain an optimized 3D resource scheduling model. Alternatively, the 3D resource scheduling model can be iteratively optimized using the device priorities of each device as the optimization objective until a preset convergence condition is met, thereby generating the optimized 3D resource scheduling model.

[0037] S105: Generate a resource scheduling scheme based on the optimized three-dimensional resource scheduling model, and perform resource scheduling on the subway system according to the resource scheduling scheme.

[0038] Specifically, in order to transform the optimized model into executable instructions and improve system response efficiency, a resource scheduling scheme needs to be generated based on the optimized three-dimensional resource scheduling model, and then the subway system's resources are scheduled according to the resource scheduling scheme. The resource scheduling scheme refers to a set of instructions including train operation scheduling, equipment maintenance scheduling, and resource configuration scheduling; resource scheduling refers to the allocation and control of subway system resources based on the resource scheduling scheme.

[0039] Regarding this step, in some possible implementations, the optimized 3D resource scheduling model can be solved to generate an initial resource scheduling instruction set, which is then coordinated to form the final resource scheduling scheme. In other possible implementations, resource scheduling instructions can be directly output based on the optimized 3D resource scheduling model and distributed to relevant execution units to complete the resource scheduling of the subway system.

[0040] In this embodiment, firstly, the first multimodal data corresponding to the subway system is acquired, and data fusion is performed based on the first multimodal data to generate a relational database. Then, the equipment priority of each device in the subway system is determined based on the relational database. Next, a three-dimensional resource scheduling model corresponding to the subway system is constructed based on spatial, temporal, and cost dimensions. Then, the three-dimensional resource scheduling model is collaboratively optimized based on the equipment priority of each device to obtain an optimized three-dimensional resource scheduling model. Finally, a resource scheduling scheme is generated based on the optimized three-dimensional resource scheduling model, and resource scheduling of the subway system is performed according to this scheme. Through the above process, the relational database generated based on the first multimodal data provides a comprehensive data foundation for determining the equipment priority of each device; the construction of a three-dimensional resource scheduling model including spatial, temporal, and cost dimensions establishes multi-dimensional optimization objectives for collaborative optimization; furthermore, the collaborative optimization of the three-dimensional resource scheduling model based on equipment priority enables the generated resource scheduling scheme to comprehensively consider multi-dimensional factors and key equipment requirements, thereby improving the accuracy of resource scheduling in the subway system.

[0041] In one embodiment, the step of "obtaining the first multimodal data corresponding to the subway system" can be further refined and may include the following steps: Acquire the equipment status data corresponding to the subway system. The equipment status data is used to characterize the real-time operating conditions and fault trends of the equipment in the subway system. Obtain passenger flow data corresponding to the subway system. Passenger flow data is used to characterize the spatiotemporal distribution and congestion situation of passenger groups in the subway system. Obtain environmental data corresponding to the subway system. The environmental data is used to characterize the parameter changes and influencing factors of environmental conditions in the subway system. Based on the equipment status data, passenger flow data, and environmental data corresponding to the subway system, the first multimodal data corresponding to the subway system is determined.

[0042] Specifically, considering the complexity of subway system operation scenarios and the heterogeneity of multi-source data, this embodiment proposes a scheme to construct the first multimodal data through hierarchical collection and fusion.

[0043] On the one hand, it is necessary to obtain the equipment status data corresponding to the subway system. The equipment status data is used to characterize the real-time operating conditions and fault trends of the equipment in the subway system. Among them, equipment status data refers to the real-time operating parameters collected by vibration sensors, temperature sensors, and current sensors deployed in key locations such as train bogies, tracks, and power supply systems; real-time operating conditions of the equipment refer to the operating status parameters of the equipment at the current moment, including but not limited to vibration frequency, temperature value, and current intensity; and fault trends of the equipment refer to the potential fault probability and evolution direction of the equipment based on the analysis of historical operating data and real-time parameters.

[0044] Regarding this step, in some possible implementations, the operating parameters of each device can be collected through an integrated system of vehicle-mounted terminals and ground terminals, and the collected raw device operating parameters can be used as device status data.

[0045] On the other hand, it is necessary to obtain passenger flow data corresponding to the subway system. Passenger flow data is used to characterize the spatiotemporal distribution and congestion status of passenger groups in the subway system. Among them, passenger flow data refers to passenger movement information collected through gate counting, video analysis, and Wi-Fi probe technology; passenger group refers to all passengers in the subway system; spatiotemporal distribution of passenger group refers to the distribution density of passengers at different time points and spatial locations; congestion status of passenger group refers to the changes in traffic efficiency caused by the degree of passenger gathering in a specific area or time period.

[0046] Regarding this step, in some possible implementations, gate counting data, congestion point data from video recognition, and anonymous mobile phone signal data from Wi-Fi probes can be combined to generate a real-time heat map of passenger flow density through multi-source fusion technology, and this heat map data can be used as passenger flow data.

[0047] On the other hand, it is necessary to obtain environmental data corresponding to the subway system. Environmental data is used to characterize the parameter changes and influencing factors of environmental conditions in the subway system. Among them, environmental data refers to parameters such as temperature, humidity, and air quality collected by meteorological sensors and environmental monitoring equipment; environmental conditions refer to the physical environmental state inside and around the subway system; parameter changes of environmental conditions refer to the fluctuations of environmental parameters over time or due to external factors; and influencing factors of environmental conditions refer to environmental variables that have a direct or indirect effect on equipment operation or passenger flow distribution.

[0048] Regarding this step, in some possible implementations, environmental monitoring equipment deployed in stations and tunnels can be used to collect parameters such as temperature and humidity in real time, and the collected raw environmental parameters can be used as environmental data.

[0049] Next, based on the equipment status data, passenger flow data, and environmental data corresponding to the subway system, the first multimodal data corresponding to the subway system is determined.

[0050] Regarding this step, in some possible implementations, equipment status data, passenger flow data, and environmental data can be aligned with timestamps and input into a multimodal fusion framework to generate first multimodal data containing equipment-passenger flow-environment association information.

[0051] In this embodiment, layered data acquisition and multi-source fusion technology are used to ensure that the first multimodal data can comprehensively reflect the equipment status, passenger flow dynamics, and environmental conditions of the subway system, providing a data foundation for subsequent construction of a related database and optimization of resource scheduling. Specifically, the synchronous acquisition and timestamp alignment of equipment status data, passenger flow data, and environmental data achieves spatiotemporal consistency of multi-source heterogeneous data, providing accurate data input for establishing the mapping relationship between equipment failures and passenger flow impacts. The construction of the multimodal fusion framework enables the system to identify equipment operation modes, passenger flow evolution patterns, and environmental parameter coupling effects based on the integrated data features, thereby providing multi-dimensional decision-making basis for the collaborative optimization of spatial, temporal, and cost dimensions in the three-dimensional resource scheduling model, ultimately improving the accuracy and dynamic adaptability of the resource scheduling scheme.

[0052] In one embodiment, the above step "data fusion based on the first multimodal data to generate an associated database" can be further refined and may include the following steps: The station-level edge computing nodes of the subway system are invoked to preprocess the first multimodal data to obtain the second multimodal data; Data fusion is performed based on the second multimodal data to generate a related database.

[0053] Specifically, considering the heterogeneity of data sources and the real-time requirements of the subway system, this embodiment proposes a scheme to construct an associated database through edge computing node preprocessing and multimodal fusion technology.

[0054] Firstly, to reduce data transmission load and improve processing efficiency, it is necessary to call the station-level edge computing nodes of the subway system to preprocess the first multimodal data to obtain the second multimodal data. The station-level edge computing nodes of the subway system refer to distributed computing units deployed at the station-level gateway, used for initial filtering and anomaly detection of the raw data; preprocessing refers to data cleaning, standardization, and timestamp alignment operations on the first multimodal data.

[0055] Regarding this step, in some possible implementations, a preset filtering algorithm can be used to remove outliers from the first multimodal data, and the filtered first multimodal data can be used as the second multimodal data; or a timestamp synchronization mechanism can be used to align the multi-source data, and the aligned first multimodal data can be used as the second multimodal data.

[0056] Furthermore, in order to establish the mapping relationship between equipment and passenger flow impact, data fusion based on the second multimodal data is required to generate a relational database. Data fusion refers to integrating equipment status data, passenger flow data, and environmental data from the second multimodal data through a multimodal fusion framework to generate the relational database.

[0057] Regarding this step, some possible implementations include using feature extraction algorithms to extract statistical, aggregated, and time-series features from the second multimodal data, and using the extracted second multimodal data as input; or using a long short-term memory network to perform predictive analysis on the second multimodal data, and using the analysis results together with the second multimodal data to generate an associated database.

[0058] In this embodiment, the first multimodal data is preprocessed by calling the station-level edge computing nodes of the subway system to obtain the second multimodal data. This process reduces the amount of data transmitted to the central node by performing preliminary filtering and anomaly detection at the data source end, and improves the timeliness of subsequent processing. A related database is generated by data fusion based on the second multimodal data. This process integrates equipment status data, passenger flow data and environmental data in the second multimodal data, establishes a mapping relationship between equipment operating status and passenger flow impact, and provides multi-dimensional data support for the resource scheduling model, thereby improving the accuracy and dynamic adaptability of resource scheduling.

[0059] In one embodiment, the above step "data fusion based on second multimodal data to generate a related database" can be further refined and may include the following steps: Data cleaning is performed on the second multimodal data to obtain cleaned second multimodal data; Feature data is obtained by extracting features from the cleaned second multimodal data; The feature data is analyzed based on a pre-defined long short-term memory network to obtain prediction results; the prediction results are used to characterize the impact of equipment failure on passenger flow prediction. Based on the prediction results and characteristic data, establish the mapping relationship between equipment and passenger flow in the subway system; Based on the mapping relationship, a related database is generated.

[0060] Specifically, considering that multi-source heterogeneous data may contain noise and missing values, and that it is necessary to establish a dynamic correlation between equipment status and passenger flow impact, this embodiment proposes a step-by-step processing scheme through data cleaning, feature extraction, time series analysis, and mapping relationship construction.

[0061] First, the second multimodal data needs to be cleaned to obtain cleaned second multimodal data. Data cleaning refers to identifying and correcting outliers, missing values, and duplicate values ​​in the second multimodal data, and standardizing the data format. The cleaned second multimodal data refers to a unified set of processed equipment status data, passenger flow data, and environmental data.

[0062] Regarding this step, in some possible implementations, a statistical detection algorithm can be used to identify outliers in the second multimodal data, fill in the missing values ​​using interpolation, and use the corrected second multimodal data as the cleaned second multimodal data; or a timestamp alignment mechanism can be used to synchronize multi-source data and use the aligned second multimodal data as the cleaned second multimodal data.

[0063] Next, feature extraction needs to be performed on the cleaned second multimodal data to obtain feature data. Feature extraction refers to extracting statistical measures, time-series features, and interaction features from the cleaned second multimodal data; feature data refers to a set of quantitative indicators that characterize equipment operation modes, passenger flow fluctuation patterns, and environmental coupling effects.

[0064] Regarding this step, some possible implementations include calculating the mean, variance, and other statistics on the cleaned second multimodal data, extracting the autocorrelation features of the time series data, and using the generated statistics and time series features as feature data; or performing dimensionality reduction on the cleaned second multimodal data using principal component analysis, and using the dimensionality-reduced data as feature data.

[0065] Furthermore, to predict the dynamic impact of equipment failure on passenger flow, it is necessary to analyze the feature data based on a pre-defined long short-term memory network to obtain prediction results; these prediction results are used to characterize the predicted impact of equipment failure on passenger flow. Long short-term memory network refers to a recurrent neural network model capable of handling temporal dependencies.

[0066] Regarding this step, in some possible implementations, the feature data can be input into a long short-term memory network, and the trained model can output the probability of equipment failure and its predictive impact on passenger flow density, and this output can be used as the prediction result; or historical failure data and real-time feature data can be combined to generate the passenger flow change trend for future periods through a long short-term memory network, and this trend can be used as the prediction result.

[0067] Then, based on the prediction results and feature data, it is necessary to establish a mapping relationship between equipment and passenger flow in the subway system. Here, passenger flow refers to the spatiotemporal distribution and movement status of passenger groups in the subway system; the mapping relationship refers to the quantitative correspondence rules between equipment failure probability, environmental parameters, and changes in passenger flow.

[0068] Regarding this step, in some possible implementations, the correlation between the prediction results and feature data can be analyzed using association rule mining algorithms to generate a mapping function of the impact of equipment failure on passenger flow, and this function can be used as the mapping relationship; or the prediction results and feature data can be input into a graph neural network to construct an association matrix between equipment nodes and passenger flow nodes, and this matrix can be used as the mapping relationship.

[0069] Finally, a related database needs to be generated based on the mapping relationship.

[0070] Regarding this step, some possible implementations include storing the mapping relationship in a relational database and creating an index to support real-time queries; this database is a relational database. Alternatively, the mapping relationship can be converted into key-value pairs and stored in a distributed caching system, which is also a relational database.

[0071] In this embodiment, outliers, missing values, and duplicate values ​​in the second multimodal data are identified and corrected through data cleaning, ensuring the data quality for subsequent feature extraction. Feature extraction extracts statistical quantities, temporal features, and interaction features from the cleaned second multimodal data, generating feature data that characterizes equipment operation modes, passenger flow fluctuation patterns, and environmental coupling effects. The feature data is analyzed using a pre-defined long short-term memory network to obtain prediction results characterizing the predictive impact of equipment failures on passenger flow, achieving temporal prediction of the dynamic correlation between equipment failures and passenger flow. Based on the prediction results and feature data, a mapping relationship between equipment and passenger flow in the subway system is established, quantifying the correspondence rules between equipment failure probabilities, environmental parameters, and passenger flow changes. A relational database is generated based on the mapping relationship, providing a high-precision, low-latency decision-making basis for resource scheduling.

[0072] In one embodiment, the above step of "determining the equipment priority of each device in the subway system based on the associated database" can be implemented based on the following formula: ; in Parameters are dynamically adjusted through a long short-term memory network. Specifically, This indicates the device's priority value; This indicates the historical frequency of failures, used to quantify the frequency of failures occurring in the equipment during its historical operation. This represents the passenger flow impact coefficient, used to quantify the potential impact of equipment failure on subway passenger flow distribution. This indicates the maintenance cost weight, used to quantify the importance of the economic costs required for equipment maintenance; , , They represent , and The corresponding weight coefficients, and satisfying Regarding this step, in some possible implementations, the parameters of each device can be calculated based on historical fault data, passenger flow data, and cost data in the associated database, using the following method: First, count the historical fault counts of each device within a preset time window, and then calculate the ratio between these counts and the total operating time of the device to obtain the parameters for each device. Secondly, based on equipment location, function type, and passenger flow data in the relational database, the potential impact of equipment failure on passenger flow indicators such as transfer channel congestion and platform congestion is analyzed, and quantified by combining the prediction results of the Long Short-Term Memory Network. Then, by integrating the spare parts procurement costs, maintenance labor costs, and downtime losses due to malfunctions from the associated database, and through cost normalization, the costs for each piece of equipment are obtained. After the calculation is completed, the system dynamically adjusts the parameters based on real-time operational status (such as peak hours or unexpected events) through the Long Short-Term Memory network. , , The value; finally, the calculated value , , and dynamically adjusted , , Substituting into the formula, the device priority of each device is calculated. In some possible implementations, future trends can be predicted using long short-term memory networks based on real-time status data in an associated database, and parameters can be updated to optimize device priorities.

[0073] In this embodiment, by constructing a quantitative formula that includes fault history frequency, passenger flow impact coefficient, and maintenance cost weight, multiple key attributes of the equipment are integrated into a single, comparable equipment priority value. This overcomes the subjectivity and ambiguity inherent in manual qualitative assessment. Furthermore, it utilizes a long short-term memory network to dynamically adjust the weight coefficients of each dimension. , , This allows equipment priority assessment to adapt to changes in real-time operational scenarios, such as automatically increasing the passenger flow impact coefficient during peak hours. The weights of these weights guide resource allocation to critical equipment that has the greatest impact on passengers. Ultimately, this equipment priority provides reliable data input for the subsequent collaborative optimization of the 3D resource scheduling model, improving the accuracy and timeliness of resource scheduling.

[0074] In one embodiment, the step of "constructing a three-dimensional resource scheduling model for the subway system based on spatial, temporal, and cost dimensions" can be further refined and may include the following steps: Based on the geographical distribution of resources and network topology of the subway system, a first model corresponding to the spatial dimension is constructed. Based on the resource scheduling time, task execution time, and resource availability time of the subway system, a second model corresponding to the time dimension is constructed. Based on the resource procurement cost, resource transportation cost, resource storage cost, and resource operation cost of the subway system, a third cost-dimensional model is constructed. Based on the first model, the second model, and the third model, a three-dimensional resource scheduling model corresponding to the subway system is constructed.

[0075] Specifically, considering that the allocation of resources in the subway system needs to take into account geographical constraints, timeliness requirements, and economy, this embodiment proposes a scheme to construct a three-dimensional resource scheduling model through hierarchical modeling and integration.

[0076] First, a first model corresponding to the spatial dimension needs to be constructed based on the geographical distribution of resources and the network topology of the subway system. The geographical distribution of resources in the subway system refers to the physical coordinates of resources such as trains, power supply systems, and maintenance equipment within the rail transit network; the network topology of the subway system refers to the connection relationships and path structure formed by stations, track sections, and transfer nodes; and the first model refers to a mathematical model that quantifies the spatial distance between resources and the constraints of transportation paths.

[0077] Regarding this step, in some possible implementations, resource coordinates can be collected through a geographic information system, network topology maps can be generated by combining graph theory algorithms, and the geographical distribution of resources and network topology can be input into a spatial optimization algorithm to output the first model.

[0078] Next, a second model corresponding to the time dimension needs to be constructed based on the resource scheduling time, task execution time, and resource availability time of the subway system. Here, the resource scheduling time of the subway system refers to the time point when resource allocation instructions are generated and issued; the task execution time of the subway system refers to the duration of actual resource usage; the resource availability time of the subway system refers to the period during which resources are idle and can be scheduled; and the second model refers to a planning model that constrains the resource task sequence based on the time axis.

[0079] Regarding this step, in some possible implementations, resource scheduling time, task execution time, and resource availability time can be input into a time window partitioning algorithm to generate discretized time segments, and a second model can be constructed through temporal logic constraints.

[0080] Then, a third cost-dimensional model needs to be constructed based on the resource procurement cost, resource transportation cost, resource storage cost, and resource operation cost of the subway system. Here, the resource procurement cost of the subway system refers to the cost of acquiring new resources; the resource transportation cost refers to the cost incurred in transferring resources geographically; the resource storage cost refers to the management and maintenance costs during resource idle periods; and the resource operation cost refers to the energy consumption and manpower expenditure during resource use. The third model is a multi-objective optimization model with the goal of minimizing costs.

[0081] Regarding this step, in some possible implementations, resource procurement costs, resource transportation costs, resource storage costs, and resource operation costs can be quantified into economic indicators, integrated into a cost function through linear weighting or the analytic hierarchy process, and a third model can be constructed based on this function.

[0082] Finally, based on the first model, the second model, and the third model, a three-dimensional resource scheduling model corresponding to the subway system is constructed.

[0083] Regarding this step, in some possible implementations, the first, second, and third models can be coupled through a multi-objective optimization framework, with the spatial network as a constraint, the time axis as a reference, and the cost function as the objective, to generate a three-dimensional resource scheduling model.

[0084] In this embodiment, a first model is constructed based on the geographical distribution and network topology of resources in the subway system, clarifying the distribution and path constraints of resources in physical space and providing a spatial feasibility basis for resource scheduling. A second model is constructed based on the resource scheduling time, task execution time, and resource availability time of the subway system, establishing task conflicts and availability windows of resources in the time series and providing a time-dimensional logical constraint for resource scheduling. A third model is constructed based on the resource procurement cost, resource transportation cost, resource storage cost, and resource operation cost of the subway system, quantifying the economic cost in the resource allocation process and providing a cost-dimensional optimization objective for resource scheduling. Finally, by integrating the first, second, and third models, a three-dimensional resource scheduling model is constructed, enabling the model to simultaneously satisfy multi-dimensional constraints of spatial path, time window, and cost-effectiveness, thereby achieving a synergistic balance between geographical accessibility, timeliness feasibility, and economic rationality in resource scheduling decisions.

[0085] In one embodiment, the step of "cooperatively optimizing the three-dimensional resource scheduling model based on the device priority of each device to obtain an optimized three-dimensional resource scheduling model" can be further refined and may include the following steps: Based on the device priority of each device, the optimization weight coefficients corresponding to the spatial dimension, time dimension and cost dimension in the three-dimensional resource scheduling model are dynamically adjusted to obtain the adjusted optimization weight coefficients; Based on the adjusted optimization weight coefficients, the three-dimensional resource scheduling model is solved through collaborative optimization to generate an optimized three-dimensional resource scheduling model.

[0086] Specifically, considering the dynamic impact of device priority on resource allocation strategies and the coupling relationship between multi-dimensional objectives, this embodiment proposes an optimization scheme based on priority weight collaborative adjustment.

[0087] Firstly, to ensure the resource scheduling model can dynamically respond to the needs of critical equipment, it is necessary to dynamically adjust the optimized weight coefficients corresponding to the spatial, temporal, and cost dimensions in the three-dimensional resource scheduling model based on the equipment priority of each device, thus obtaining the adjusted optimized weight coefficients. Adjusting the optimized weight coefficients corresponding to the spatial, temporal, and cost dimensions in the three-dimensional resource scheduling model refers to updating the weights of spatial path constraints, time window constraints, and cost objective functions in real time based on the quantified values ​​of equipment priority (such as historical failure frequency, passenger flow impact coefficient, and maintenance cost weight). The adjusted optimized weight coefficients refer to the set of weight parameters for the spatial, temporal, and cost dimensions after dynamic prioritization. Regarding this step, in some possible implementations, the device priority of each device can be mapped to the weight coefficients of the spatial, temporal, and cost dimensions in the three-dimensional resource scheduling model through a preset weight allocation algorithm (such as linear weighting), and the mapped weight coefficients can be used as the adjusted optimized weight coefficients; or a reinforcement learning model can be used with device priority as the state input, output the adjusted value of the optimized weight coefficients, and the adjusted value can be superimposed with the original weight coefficients to generate the adjusted optimized weight coefficients.

[0088] Furthermore, based on the adjusted optimization weight coefficients, the three-dimensional resource scheduling model is collaboratively optimized to generate an optimized three-dimensional resource scheduling model. Collaborative optimization refers to using the adjusted optimization weight coefficients as constraints and employing a multi-objective optimization algorithm (such as a genetic algorithm) to jointly solve for resource allocation objectives across spatial, temporal, and cost dimensions.

[0089] Regarding this step, in some possible implementations, the adjusted optimization weight coefficients can be input into the three-dimensional resource scheduling model, and a Pareto optimal solution set can be generated by the non-dominated sorting genetic algorithm (NSGA-II). The optimal solution in the solution set can then be used as the optimized three-dimensional resource scheduling model. Alternatively, the Lagrange relaxation method can be used to transform the adjusted optimization weight coefficients into constraints, and the three-dimensional resource scheduling model can be iteratively solved until the objective function converges. The converged model can then be used as the optimized three-dimensional resource scheduling model.

[0090] In this embodiment, the weight coefficients of the equipment priority are dynamically adjusted to optimize the system. This allows the target weights of the spatial, temporal, and cost dimensions in the three-dimensional resource scheduling model to change in real time according to the criticality of the equipment, thereby directly linking the resource allocation strategy with the demand for critical equipment. Through collaborative optimization, the system integrates multi-dimensional constraints of spatial paths, time windows, and cost targets. This ensures that the optimized three-dimensional resource scheduling model meets the timeliness and spatial accessibility requirements of equipment priority while guaranteeing the overall economy of the system through the weight constraints of the cost dimension, ultimately achieving a balance between system reliability and economy.

[0091] In one embodiment, the step of "generating a resource scheduling scheme based on the optimized three-dimensional resource scheduling model" can be further refined and may include the following steps: The optimized three-dimensional resource scheduling model is solved to generate an initial resource scheduling instruction set; wherein, the initial resource scheduling instruction set includes at least one of train operation scheduling instructions, equipment maintenance scheduling instructions, and resource configuration scheduling instructions; The station-level edge computing nodes of the subway system are invoked to globally coordinate the initial resource scheduling instruction set based on a preset distributed consensus algorithm, resulting in a coordinated resource scheduling instruction set. Based on the coordinated resource scheduling instruction set, a resource scheduling scheme is generated.

[0092] Specifically, considering that the resource scheduling scheme needs to meet multi-dimensional constraints and ensure consistency among distributed nodes, this embodiment proposes a step-by-step processing scheme through model solving, global coordination, and instruction generation.

[0093] First, to transform the optimized 3D resource scheduling model into executable scheduling instructions, it is necessary to solve the optimized 3D resource scheduling model to generate an initial resource scheduling instruction set. This initial resource scheduling instruction set includes at least one of the following: train operation scheduling instructions, equipment maintenance scheduling instructions, and resource allocation scheduling instructions. Train operation scheduling instructions refer to control instructions that adjust train departure intervals, operating routes, and station dwell times; equipment maintenance scheduling instructions refer to scheduling instructions that allocate maintenance resources, set maintenance periods and priorities; and resource allocation scheduling instructions refer to instructions that allocate system resources such as power, ventilation, and air conditioning.

[0094] Regarding this step, some possible implementations include using a multi-objective optimization algorithm to solve the optimized 3D resource scheduling model and generating an initial resource scheduling instruction set containing train operation scheduling instructions, equipment maintenance scheduling instructions, and resource configuration scheduling instructions; or using an integer programming method to discretize and solve the optimized 3D resource scheduling model and using the solution result as the initial resource scheduling instruction set.

[0095] Next, the station-level edge computing nodes of the subway system are invoked to globally coordinate the initial resource scheduling instruction set based on a preset distributed consensus algorithm, resulting in a coordinated resource scheduling instruction set. Here, the station-level edge computing nodes of the subway system refer to distributed computing units deployed at the station-level gateway, used for instruction coordination and conflict resolution; the distributed consensus algorithm refers to a consensus mechanism that coordinates data updates through master nodes and synchronizes data with slave nodes in real time; and global coordination ensures that all edge nodes have a consistent understanding of the resource scheduling instruction set.

[0096] Regarding this step, some possible implementations include calling the station-level edge computing nodes of the subway system to perform conflict detection and priority sorting on the train operation scheduling instructions, equipment maintenance scheduling instructions, and resource configuration scheduling instructions in the initial resource scheduling instruction set through a preset distributed consensus algorithm, and generating a coordinated resource scheduling instruction set; or using the station-level edge computing nodes of the subway system to perform consistency verification on the initial resource scheduling instruction set based on a preset distributed consensus algorithm, and using the verified initial resource scheduling instruction set as the coordinated resource scheduling instruction set.

[0097] Finally, a resource scheduling scheme is generated based on the coordinated resource scheduling instruction set.

[0098] Regarding this step, some possible implementations include encapsulating the coordinated resource scheduling instruction set into a standardized scheduling message to generate a resource scheduling scheme; or converting the coordinated resource scheduling instruction set into a resource scheduling scheme that can be directly called by the subway system execution unit through preset instruction mapping rules.

[0099] Understandably, since the resource scheduling scheme is generated based on the coordinated resource scheduling instruction set, the specific implementation process of "scheduling resources for the subway system according to the resource scheduling scheme" can be expressed as follows: the train operation scheduling instructions in the resource scheduling scheme are sent to the train operation control system, the equipment maintenance scheduling instructions are sent to the maintenance resource management system, and the resource configuration scheduling instructions are sent to the power supply system and the ventilation and air conditioning system, thereby realizing dynamic resource scheduling of the subway system.

[0100] In this embodiment, an initial resource scheduling instruction set is generated by solving the optimized three-dimensional resource scheduling model. The optimization objectives of spatial, temporal, and cost dimensions are transformed into specific train operation scheduling instructions, equipment maintenance scheduling instructions, and resource configuration scheduling instructions, ensuring the compliance of the resource scheduling scheme with multi-dimensional constraints. By calling the station-level edge computing nodes of the metro system and globally coordinating the initial resource scheduling instruction set based on a preset distributed consensus algorithm, the scheduling inconsistency problem caused by data synchronization delays or instruction conflicts between distributed nodes is solved, ensuring that all nodes reach a consensus on the execution status of train operation scheduling instructions, equipment maintenance scheduling instructions, and resource configuration scheduling instructions. A resource scheduling scheme is generated based on the coordinated resource scheduling instruction set, and the globally coordinated instruction set is converted into standardized instructions that can be directly called by various execution units of the metro system. This realizes closed-loop control of the resource scheduling scheme from model to execution. Finally, through the synergistic effect of the above steps, the accuracy and response efficiency of resource allocation in complex operating scenarios of the metro system are improved.

[0101] The following will combine Figure 2 This application provides a detailed description of the resource scheduling device 800 for a subway system, which corresponds to the resource scheduling method for the subway system described above. Specifically, the resource scheduling device 800 may include an acquisition module 810, a determination module 820, a construction module 830, an optimization module 840, and a scheduling module 850, as detailed below: The acquisition module 810 is used to acquire the first multimodal data corresponding to the subway system, and perform data fusion based on the first multimodal data to generate a related database; Module 820 is used to determine the equipment priority of each device in the subway system based on the associated database. Module 830 is used to construct a three-dimensional resource scheduling model for the subway system based on spatial, temporal, and cost dimensions. The optimization module 840 is used to perform collaborative optimization of the three-dimensional resource scheduling model based on the device priority of each device, so as to obtain the optimized three-dimensional resource scheduling model. The scheduling module 850 is used to generate a resource scheduling plan based on the optimized three-dimensional resource scheduling model, and to perform resource scheduling on the subway system according to the resource scheduling plan.

[0102] Optionally, in some embodiments, the acquisition module 810 can be used to: Acquire the equipment status data corresponding to the subway system. The equipment status data is used to characterize the real-time operating conditions and fault trends of the equipment in the subway system. Obtain passenger flow data corresponding to the subway system. Passenger flow data is used to characterize the spatiotemporal distribution and congestion situation of passenger groups in the subway system. Obtain environmental data corresponding to the subway system. The environmental data is used to characterize the parameter changes and influencing factors of environmental conditions in the subway system. Based on the equipment status data, passenger flow data, and environmental data corresponding to the subway system, the first multimodal data corresponding to the subway system is determined.

[0103] Optionally, in some embodiments, the acquisition module 810 can be used to: The station-level edge computing nodes of the subway system are invoked to preprocess the first multimodal data to obtain the second multimodal data; Data fusion is performed based on the second multimodal data to generate a related database.

[0104] Optionally, in some embodiments, the acquisition module 810 can be used to: Data cleaning is performed on the second multimodal data to obtain cleaned second multimodal data; Feature data is obtained by extracting features from the cleaned second multimodal data; The feature data is analyzed based on a pre-defined long short-term memory network to obtain prediction results; the prediction results are used to characterize the impact of equipment failure on passenger flow prediction. Based on the prediction results and characteristic data, establish the mapping relationship between equipment and passenger flow in the subway system; Based on the mapping relationship, a related database is generated.

[0105] Optionally, in some embodiments, the building module 830 can be used to: Based on the geographical distribution of resources and network topology of the subway system, a first model corresponding to the spatial dimension is constructed. Based on the resource scheduling time, task execution time, and resource availability time of the subway system, a second model corresponding to the time dimension is constructed. Based on the resource procurement cost, resource transportation cost, resource storage cost, and resource operation cost of the subway system, a third cost-dimensional model is constructed. Based on the first model, the second model, and the third model, a three-dimensional resource scheduling model corresponding to the subway system is constructed.

[0106] Optionally, in some embodiments, the optimization module 840 can be used to: Based on the device priority of each device, the optimization weight coefficients corresponding to the spatial dimension, time dimension and cost dimension in the three-dimensional resource scheduling model are dynamically adjusted to obtain the adjusted optimization weight coefficients; Based on the adjusted optimization weight coefficients, the three-dimensional resource scheduling model is solved through collaborative optimization to generate an optimized three-dimensional resource scheduling model.

[0107] Optionally, in some embodiments, the scheduling module 850 can be used to: The optimized three-dimensional resource scheduling model is solved to generate an initial resource scheduling instruction set; wherein, the initial resource scheduling instruction set includes at least one of train operation scheduling instructions, equipment maintenance scheduling instructions, and resource configuration scheduling instructions; The station-level edge computing nodes of the subway system are invoked to globally coordinate the initial resource scheduling instruction set based on a preset distributed consensus algorithm, resulting in a coordinated resource scheduling instruction set. Based on the coordinated resource scheduling instruction set, a resource scheduling scheme is generated.

[0108] For the effects achievable in this embodiment, please refer to the relevant embodiments of the resource scheduling method for the subway system described above, which will not be repeated here.

[0109] Figure 3 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 3 As shown, the electronic device may include: a processor 1301, a communication interface 1302, a memory 1303, and a communication bus 1304. The processor 1301, communication interface 1302, and memory 1303 communicate with each other via the communication bus 1304. The processor 1301 can call a computer program stored in the memory 1303 to execute steps of a resource scheduling method for the subway system, such as: Obtain the first multimodal data corresponding to the subway system, and perform data fusion based on the first multimodal data to generate a related database; Determine the equipment priority of each device in the subway system based on the associated database; Based on spatial, temporal, and cost dimensions, a three-dimensional resource scheduling model for the subway system is constructed. Based on the device priority of each device, the 3D resource scheduling model is collaboratively optimized to obtain the optimized 3D resource scheduling model. A resource scheduling scheme is generated based on the optimized 3D resource scheduling model, and the subway system is then scheduled according to the resource scheduling scheme.

[0110] Furthermore, when the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0111] On the other hand, embodiments of this application also provide a non-transitory computer-readable storage medium storing a computer program. The computer program is used to cause a processor to execute the steps of the methods provided in the above embodiments, including, for example: Obtain the first multimodal data corresponding to the subway system, and perform data fusion based on the first multimodal data to generate a related database; Determine the equipment priority of each device in the subway system based on the associated database; Based on spatial, temporal, and cost dimensions, a three-dimensional resource scheduling model for the subway system is constructed. Based on the device priority of each device, the 3D resource scheduling model is collaboratively optimized to obtain the optimized 3D resource scheduling model. A resource scheduling scheme is generated based on the optimized 3D resource scheduling model, and the subway system is then scheduled according to the resource scheduling scheme.

[0112] Non-transitory computer-readable storage media can be any available medium or data storage device that can be accessed by a processor, including but not limited to magnetic storage (e.g., floppy disks, hard disks, magnetic tapes, magneto-optical disks (MOs), etc.), optical storage (e.g., CDs, DVDs, BDs, HVDs, etc.), and semiconductor storage (e.g., ROMs, EPROMs, EEPROMs, non-volatile memory (NAND flash), solid-state drives (SSDs)).

[0113] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0114] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments or some parts of embodiments.

[0115] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A resource scheduling method for a subway system, characterized in that, include: Obtain the first multimodal data corresponding to the subway system, and perform data fusion based on the first multimodal data to generate a related database; The device priority of each device in the subway system is determined based on the associated database. Based on spatial, temporal, and cost dimensions, a three-dimensional resource scheduling model corresponding to the subway system is constructed. Based on the device priority of each device, the three-dimensional resource scheduling model is collaboratively optimized to obtain the optimized three-dimensional resource scheduling model. A resource scheduling scheme is generated based on the optimized three-dimensional resource scheduling model, and the subway system is then scheduled according to the resource scheduling scheme.

2. The method according to claim 1, characterized in that, The acquisition of the first multimodal data corresponding to the subway system includes: Obtain the equipment status data corresponding to the subway system. The equipment status data is used to characterize the real-time operating conditions and fault trends of the equipment in the subway system. Obtain passenger flow data corresponding to the subway system, and the passenger flow data is used to characterize the spatiotemporal distribution and congestion situation of the passenger group in the subway system; Obtain environmental data corresponding to the subway system, and the environmental data is used to characterize the parameter changes and influencing factors of environmental conditions in the subway system; Based on the equipment status data, passenger flow data, and environmental data corresponding to the subway system, the first multimodal data corresponding to the subway system is determined.

3. The method according to claim 1, characterized in that, The step of performing data fusion based on the first multimodal data to generate a related database includes: The station-level edge computing node of the subway system is invoked to preprocess the first multimodal data to obtain the second multimodal data; Data fusion is performed based on the second multimodal data to generate an associated database.

4. The method according to claim 3, characterized in that, The step of data fusion based on the second multimodal data to generate a related database includes: The second multimodal data is cleaned to obtain cleaned second multimodal data; Feature data is obtained by extracting features from the cleaned second multimodal data. The feature data is analyzed based on a pre-defined long short-term memory network to obtain prediction results; wherein, the prediction results are used to characterize the impact of equipment failure on passenger flow prediction. Based on the prediction results and the feature data, a mapping relationship between equipment and passenger flow in the subway system is established; Based on the mapping relationship, an associated database is generated.

5. The method according to claim 1, characterized in that, The construction of a three-dimensional resource scheduling model for the subway system based on spatial, temporal, and cost dimensions includes: Based on the resource geographic location distribution and network topology of the subway system, a first model corresponding to the spatial dimension is constructed. Based on the resource scheduling time, task execution time, and resource availability time of the subway system, a second model corresponding to the time dimension is constructed. Based on the resource procurement cost, resource transportation cost, resource storage cost, and resource operation cost of the subway system, a third cost-dimensional model is constructed. Based on the first model, the second model, and the third model, a three-dimensional resource scheduling model corresponding to the subway system is constructed.

6. The method according to claim 1, characterized in that, The step of collaboratively optimizing the three-dimensional resource scheduling model based on the device priorities of each device to obtain an optimized three-dimensional resource scheduling model includes: Based on the device priority of each device, the optimization weight coefficients corresponding to the spatial dimension, time dimension and cost dimension in the three-dimensional resource scheduling model are dynamically adjusted to obtain the adjusted optimization weight coefficients. Based on the adjusted optimization weight coefficients, the three-dimensional resource scheduling model is solved through collaborative optimization to generate an optimized three-dimensional resource scheduling model.

7. The method according to claim 1, characterized in that, The step of generating a resource scheduling scheme based on the optimized three-dimensional resource scheduling model includes: The optimized three-dimensional resource scheduling model is solved to generate an initial resource scheduling instruction set; wherein, the initial resource scheduling instruction set includes at least one of train operation scheduling instructions, equipment maintenance scheduling instructions, and resource configuration scheduling instructions; The station-level edge computing nodes of the subway system are invoked to globally coordinate the initial resource scheduling instruction set based on a preset distributed consensus algorithm, thereby obtaining a coordinated resource scheduling instruction set. Based on the coordinated resource scheduling instruction set, a resource scheduling scheme is generated.

8. A resource scheduling device for a subway system, characterized in that, include: The acquisition module is used to acquire the first multimodal data corresponding to the subway system, and perform data fusion based on the first multimodal data to generate a related database; The determination module is used to determine the equipment priority of each device in the subway system based on the associated database; The construction module is used to construct a three-dimensional resource scheduling model corresponding to the subway system based on spatial, temporal, and cost dimensions. The optimization module is used to collaboratively optimize the three-dimensional resource scheduling model based on the device priority of each device to obtain an optimized three-dimensional resource scheduling model. The scheduling module is used to generate a resource scheduling scheme based on the optimized three-dimensional resource scheduling model, and to perform resource scheduling on the subway system according to the resource scheduling scheme.

9. An electronic device comprising a processor and a memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the resource scheduling method for the subway system according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the resource scheduling method for the subway system as described in any one of claims 1 to 7.

Images