Resource scheduling method, device, equipment and medium based on heterogeneous transportation network

Abstract

The present disclosure provides a resource scheduling method, apparatus, device, and medium based on a heterogeneous transportation network, comprising: obtaining traffic flow data between a resource scheduling start location and a resource scheduling end location; extracting traffic spatiotemporal features from the traffic flow data; obtaining high-level semantic features in a resource heterogeneous graph; determining multiple candidate scheduling strategies corresponding to the transportation of mobile resources from the resource scheduling start location to the resource scheduling end location based on the traffic spatiotemporal features and the high-level semantic features; and determining, from the multiple candidate scheduling strategies, a target scheduling strategy corresponding to the transportation of mobile resources from the resource scheduling start location to the resource scheduling end location based on the scheduling loss value corresponding to each candidate scheduling strategy. Thus, by analyzing the relationships and connection paths between different transportation nodes, the accuracy of resource scheduling is effectively improved.

Description

Technical Field

[0001] The embodiments of the present disclosure relate to the field of smart transportation technology, and in particular, to a resource scheduling method, apparatus, device, and medium applicable to a heterogeneous transportation network. Background Art

[0002] With the accelerating pace of globalization, modern transportation networks are becoming increasingly complex. Transportation hubs encompass not only land transportation (such as roads and railways) but also multiple modes of transport, including air and water transport. The interweaving and interconnected nature of these various transportation networks makes the efficient scheduling and rational allocation of resources a pressing challenge. Especially in the event of emergencies, urgent transportation, and complex weather conditions, effectively coordinating multiple modes of transportation to ensure that resources reach their destinations quickly, safely, and cost-effectively has become a key research topic in the field of smart transportation and logistics.

[0003] In the related art, the existing scheduling method mainly adopts a regularized scheduling method, which implements resource transmission scheduling by formulating certain scheduling rules.

[0004] However, using the existing method, the resource scheduling accuracy is not high. Summary of the Invention

[0005] The embodiments described herein provide a method, apparatus, device, and medium for resource scheduling based on a heterogeneous transportation network, which overcome the above-mentioned problems.

[0006] In a first aspect, according to the present disclosure, a resource scheduling method based on a heterogeneous transportation network is provided, comprising:

[0007] Acquire traffic flow data between a resource scheduling starting position and a resource scheduling ending position, wherein the resource scheduling starting position is a starting transportation position of the mobile resource and the resource scheduling ending position is a target transportation position of the mobile resource;

[0008] Extracting the spatiotemporal characteristics of traffic between the resource scheduling starting position and the resource scheduling ending position from the traffic flow data between the resource scheduling starting position and the resource scheduling ending position;

[0009] Obtaining high-level semantic features in a resource heterogeneous graph, wherein the resource heterogeneous graph includes: a plurality of resource nodes and connection edges between different resource nodes, wherein the resource nodes have corresponding node types, wherein the node types are used to describe different types of resources, resource transportation equipment, and resource transportation methods, and wherein the connection edges have corresponding relationship types, wherein the relationship types are used to describe transportation equipment transportation resources and transportation equipment transportation methods;

[0010] Based on the traffic spatiotemporal characteristics and the high-level semantic characteristics, determining a plurality of candidate scheduling strategies corresponding to the transportation of the mobile resource from the resource scheduling starting location to the resource scheduling ending location, the candidate scheduling strategies including: resource scheduling duration, resource scheduling cost, and resource scheduling mode;

[0011] From the plurality of candidate scheduling strategies, based on the scheduling loss value corresponding to each candidate scheduling strategy, a target scheduling strategy corresponding to the transportation of the mobile resource from the resource scheduling starting position to the resource scheduling ending position is determined.

[0012] In a second aspect, according to the present disclosure, a resource scheduling device based on a heterogeneous transportation network is provided, comprising:

[0013] A first acquisition module is configured to acquire traffic flow data between a resource scheduling start position and a resource scheduling end position, wherein the resource scheduling start position is a starting transportation position of a mobile resource and the resource scheduling end position is a target transportation position of the mobile resource;

[0014] an extraction module, configured to extract, from the traffic flow data between the resource scheduling starting position and the resource scheduling ending position, the spatiotemporal characteristics of traffic between the resource scheduling starting position and the resource scheduling ending position;

[0015] A second acquisition module is configured to acquire high-level semantic features in a resource heterogeneous graph, wherein the resource heterogeneous graph includes: a plurality of resource nodes and connection edges between different resource nodes, wherein the resource nodes have corresponding node types, wherein the node types are used to describe different types of resources, resource transportation equipment, and resource transportation methods, and wherein the connection edges have corresponding relationship types, wherein the relationship types are used to describe transportation equipment transportation resources and transportation equipment transportation methods;

[0016] A first determination module is configured to determine, based on the traffic spatiotemporal characteristics and the high-level semantic characteristics, a plurality of candidate scheduling strategies corresponding to the transportation of the mobile resource from the resource scheduling starting location to the resource scheduling ending location, the candidate scheduling strategies including: resource scheduling duration, resource scheduling cost, and resource scheduling mode;

[0017] The second determination module is used to determine the target scheduling strategy corresponding to the transportation of the mobile resource from the resource scheduling starting position to the resource scheduling ending position from the multiple candidate scheduling strategies based on the scheduling loss value corresponding to each candidate scheduling strategy.

[0018] In a third aspect, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the steps of the resource scheduling method based on a heterogeneous transportation network in any of the above embodiments are implemented.

[0019] In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the resource scheduling method based on a heterogeneous transportation network in any of the above embodiments are implemented.

[0020] The embodiment of the present application provides a resource scheduling method based on a heterogeneous transportation network, which obtains traffic flow data between a resource scheduling start position and a resource scheduling end position, where the resource scheduling start position is the starting transportation position of the mobile resource, and the resource scheduling end position is the target transportation position of the mobile resource; extracts the traffic spatiotemporal characteristics between the resource scheduling start position and the resource scheduling end position from the traffic flow data between the resource scheduling start position and the resource scheduling end position; obtains high-level semantic features in a resource heterogeneous graph, where the resource heterogeneous graph includes: multiple resource nodes and connecting edges between different resource nodes, where the resource nodes correspond to node types, which are used to describe different types of resources, resource transportation equipment, and resource transportation methods, and where the connecting edges correspond to relationship types, which are used to describe transportation equipment transportation resources and transportation equipment transportation methods; based on the traffic spatiotemporal features and high-level semantic features, determines multiple candidate scheduling strategies corresponding to the transportation of the mobile resource from the resource scheduling start position to the resource scheduling end position, where the candidate scheduling strategies include: resource scheduling duration, resource scheduling cost, and resource scheduling method; and determines, from the multiple candidate scheduling strategies, based on the scheduling loss value corresponding to each candidate scheduling strategy, a target scheduling strategy corresponding to the transportation of the mobile resource from the resource scheduling start position to the resource scheduling end position. In this way, by analyzing the relationship and connection paths between different transportation nodes, the accuracy of resource scheduling can be effectively improved.

[0021] The above description is only an overview of the technical solutions of the embodiments of the present application. In order to more clearly understand the technical means of the embodiments of the present application, they can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the embodiments of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments will be briefly described below. It should be noted that the drawings described below only relate to some embodiments of the present disclosure and are not intended to limit the present disclosure.

[0023] Figure 1 This is a flow chart of a resource scheduling method based on a heterogeneous transportation network provided by the present disclosure.

[0024] Figure 2 This is a structural diagram of a resource scheduling device based on a heterogeneous transportation network provided by the present disclosure.

[0025] Figure 3 It is a structural diagram of a computer device provided by the present disclosure.

[0026] It should be noted that the elements in the drawings are schematic and not drawn to scale. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solutions and advantages of the embodiments of the present disclosure more clear, the technical solutions of the embodiments of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the described embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without creative work also fall within the scope of protection of the present disclosure.

[0028] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present disclosure belongs. It will be further understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the specification and the relevant art, and will not be interpreted in an idealized or overly formal manner unless otherwise explicitly defined herein. As used herein, a statement that two or more parts are "connected" or "coupled" together shall mean that the parts are joined together either directly or through one or more intermediate components.

[0029] References to "embodiments" herein mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase "embodiment" in various places in the specification does not necessarily refer to the same embodiment, nor does it necessarily refer to independent or alternative embodiments that are mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0030] The term "and / or" in this document simply describes an association between related objects, indicating that three possible relationships exist. For example, "A and / or B" can mean: A exists, A and B exist simultaneously, and B exists. Additionally, the character " / " in this document generally indicates that the related objects are in an "or" relationship. Terms such as "first" and "second" are used solely to distinguish one component (or portion of a component) from another component (or portion of a component).

[0031] In the description of this application, unless otherwise specified, "plurality" means more than two (including two), and similarly, "multiple groups" means more than two (including two).

[0032] In intelligent transportation systems, how to address these challenges and improve resource scheduling flexibility has become a focus. On the one hand, the advancement of smart cities and smart transportation has enabled more comprehensive data collection in cities, including traffic flow monitoring, real-time weather forecasts, and traffic accident information. On the other hand, with the maturity of technologies such as deep learning and graph neural networks (GNNs), it has become possible to optimize resource scheduling by leveraging multi-level, multi-dimensional, heterogeneous graph data in transportation networks. By using intelligent algorithms to analyze the relationships and connection paths between different transportation nodes, the accuracy and real-time performance of resource scheduling can be significantly improved.

[0033] Rapid resource scheduling aims to infer resource transportation methods based on known departure and destination points and the resources to be transported. This implementation, based on qualitative transportation network representation learning, fully considers how to complement the traffic map and provide complementary scheduling information, thereby shortening the scheduling method implementation and selecting the optimal scheduling method.

[0034] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0035] Figure 1 This is a flow chart of a resource scheduling method based on a heterogeneous transportation network provided by an embodiment of the present disclosure. Figure 1 As shown in Figure 2, the specific process of the resource scheduling method based on heterogeneous transportation networks includes:

[0036] S110: Obtain traffic flow data between a resource scheduling starting location and a resource scheduling ending location.

[0037] The resource scheduling starting location is the starting transportation location of the mobile resource, and the resource scheduling ending location is the target transportation location of the mobile resource. Mobile resources are resources that can be transported over long distances in real life, such as fresh food, smart devices, pets, and daily necessities.

[0038] Traffic flow data may include, but is not limited to, communication flow, traffic capacity, traffic lights, time points of passing through specific locations, special factors such as whether there is construction on land, weather conditions, width of land, etc.

[0039] Specifically, the traffic flow data between the resource scheduling start position and the resource scheduling end position can be displayed in the form of a traffic flow map, which can be obtained from a public data website and pre-processed accordingly.

[0040] For example, a traffic flow graph consists of nodes and edges. Nodes represent the traffic status at a specific location, and edges represent the connectivity of traffic routes, which can capture changes over time. Nodes have corresponding node features. Each location node vi , has a feature vector at each time step t Includes: Traffic flow Traffic capacity Traffic light signal status The time of a specific location (e.g., whether passing through a closed road) road construction Weather conditions Road width In this way, the node feature of each node can be expressed as Two nodes v i and v j The edges between them represent their traffic connection relationship (such as roads, railways, etc.), that is, the characteristics of the edges, that is, the characteristics of the edges may include: distance d ij , driving time t ij , so the edge feature can be expressed as: e ij =[d ij ,t ij ].

[0041] S120 , extracting spatiotemporal traffic features between the resource scheduling starting position and the resource scheduling ending position from traffic flow data between the resource scheduling starting position and the resource scheduling ending position.

[0042] Among them, the spatiotemporal encoder jointly preserves the contextual information on the traffic flow map and jointly models the traffic temporal patterns at different time steps and the geographical patterns of spatial regions.

[0043] In some embodiments, extracting the spatiotemporal characteristics of traffic between the resource scheduling start location and the resource scheduling end location from the traffic flow data between the resource scheduling start location and the resource scheduling end location includes:

[0044] From the traffic flow data between the resource scheduling starting position and the resource scheduling ending position, the traffic spatial characteristics between the resource scheduling starting position and the resource scheduling ending position are extracted; from the traffic flow data between the resource scheduling starting position and the resource scheduling ending position, the traffic time characteristics between the resource scheduling starting position and the resource scheduling ending position are extracted; the traffic spatial characteristics and traffic time characteristics are integrated to obtain the traffic spatiotemporal characteristics between the resource scheduling starting position and the resource scheduling ending position.

[0045] Among them, traffic space characteristics may include: horizontal traffic space (such as walkways), vertical traffic space (such as stairs, elevators, escalators) and traffic hub space, etc. Traffic time characteristics may include: traffic duration, traffic time, etc.

[0046] When integrating traffic space characteristics and traffic time characteristics, the traffic space characteristics and traffic time characteristics can be integrated in equal proportion, so as to effectively determine the traffic space-time characteristics between the starting position and the ending position of resource scheduling.

[0047] In some embodiments, extracting traffic spatial features between the resource scheduling start location and the resource scheduling end location from traffic flow data between the resource scheduling start location and the resource scheduling end location includes:

[0048] A graph convolutional network is used to extract the spatial representation information corresponding to the traffic flow data between the resource scheduling start position and the resource scheduling end position; based on the spatial representation information, the traffic spatial characteristics between the resource scheduling start position and the resource scheduling end position are determined.

[0049] Traffic spatial feature extraction is mainly achieved through graph convolutional networks (GCNs), which can capture local spatial relationships in graph structures. Graph convolution updates the node representation by summarizing the neighboring node information of each node.

[0050] For example, given a node v i and its neighbor node set N(i), the spatial feature update formula of the node in the lth layer is: in, represents the feature representation of node v at layer l+1, Represents the weight matrix, and σ represents the activation function, such as ReLU. The representation of each node is updated by the weighted summation of the features of its neighboring nodes (i.e., convolution operation) and the linear transformation of the node's own features to obtain the traffic space feature

[0051] In some embodiments, extracting the traffic time characteristics between the resource scheduling start location and the resource scheduling end location from the traffic flow data between the resource scheduling start location and the resource scheduling end location includes:

[0052] A long short-term memory network is used to extract the time representation information corresponding to the traffic flow data between the resource scheduling starting position and the resource scheduling ending position; based on the time representation information, the traffic time characteristics between the resource scheduling starting position and the resource scheduling ending position are determined.

[0053] Since traffic data is a time series, this embodiment uses LSTM (Long Short-Term Memory) to extract time features to capture temporal change trends.

[0054] For example, in LSTM, given a node v iCharacteristics at time t The update formula of LSTM is: The forget gate controls whether the state information of the previous moment is retained. Through the LSTM network, the dynamic trend of traffic flow changes over time can be captured; It is the input gate, which controls the impact of the current input on the cell state; is the candidate cell state, representing the new information at the current moment; is the cell state, storage node v i Long-term memory at time t; is the output gate, which controls the output at the current moment; ⊙ is the element-by-element multiplication; W is the corresponding matrix; b is the bias term. From the above, we can get the traffic time characteristics

[0055] S130: Obtain high-level semantic features in the resource heterogeneous graph.

[0056] Among them, the resource heterogeneous graph includes: multiple resource nodes and connecting edges between different resource nodes. Resource nodes correspond to node types, and node types are used to describe different types of resources, resource transportation equipment, and resource transportation methods. Connecting edges correspond to relationship types, and relationship types are used to describe transportation equipment transportation resources and transportation equipment transportation methods.

[0057] When constructing a resource heterogeneous graph, different types of resources, resource transportation vehicles (i.e., resource transportation equipment), and resource delivery methods (i.e., resource transportation methods) can be regarded as different types of nodes, and the relationships between resources that can be transported by different types of transportation vehicles (i.e., transportation equipment transportation resources) and the ways in which different transportation vehicles can deliver resources (i.e., transportation equipment transportation methods) can be regarded as edges.

[0058] The resource heterogeneous graph can be formally defined as: G = (V, E, A, R). V: A set of nodes, including nodes of different types, such as V = {v1, v2, ..., v n E: A set of edges, representing the relationship or connection between different nodes. Edges represent the relationship between nodes. Edges in heterogeneous graphs can be of multiple types, including transportation resources and transportation methods. A: A set of node types. Each node v i All belong to a node type A, namely A(v i )∈A, node type: resource type, available transportation and transportation delivery method. Each node type represents an entity category. The set of all nodes is the node set V. R: the set of relationship types. Each edge e ij All belong to a certain relationship type R, namely R(e ij)∈R. In a resource heterogeneous graph, different nodes have different semantic information.

[0059] The resource heterogeneous graph contains many different types of nodes (resources, hubs, transportation tools, etc.) and edges (transportation relationships, scheduling relationships, etc.). By conveying the flow of resources in the system, scheduling priorities, and complex relationships, we can obtain high-level semantic features of resource flow.

[0060] In some embodiments, obtaining high-level semantic features in a resource heterogeneous graph includes:

[0061] A heterogeneous graph neural network is used to perform representation learning on the resource heterogeneous graph to capture the structural information and semantic information in the resource heterogeneous graph; based on the structural information and semantic information in the resource heterogeneous graph, the high-level semantic features in the resource heterogeneous graph are determined.

[0062] Among them, heterogeneous graph neural networks are used for representation learning to capture the structural and semantic information in the resource heterogeneous graph. Commonly used heterogeneous graph neural networks include: Heterogeneous Graph Convolutional Network (HGCN) and MetaPath Embedding (MPE).

[0063] In a resource-heterogeneous graph, a metapath is a path defined by a series of node and edge types. Metapaths can capture complex relationships between different node types. For example, in a transportation network, a metapath might be: fresh produce - airplane - air freight. By defining metapaths, complex queries and pattern capture can be designed on the resource-heterogeneous graph, enabling further learning of node embedding representations.

[0064] When dealing with the diversity of nodes and edges, heterogeneous graph convolutional networks design different convolution kernels for different types of nodes and edges. For example, for node v i , its update can be expressed as: Nr(v) represents the set of neighbor nodes connected to node v under relationship r; is the transformation matrix of the corresponding relationship r; σ is the activation function (such as ReLU).

[0065] S140 . Determine multiple candidate scheduling strategies corresponding to the transportation of mobile resources from the resource scheduling starting location to the resource scheduling ending location based on traffic spatiotemporal characteristics and high-level semantic characteristics.

[0066] The candidate scheduling strategies include: resource scheduling duration, resource scheduling cost, and resource scheduling method. Accordingly, each candidate scheduling strategy includes a set of resource scheduling duration, resource scheduling cost, and resource scheduling method (such as air transportation, road transportation, and rail transportation).

[0067] In some embodiments, based on the spatiotemporal characteristics of traffic and high-level semantic features, multiple candidate scheduling strategies corresponding to transporting mobile resources from a resource scheduling starting location to a resource scheduling ending location are determined, including:

[0068] By integrating traffic spatiotemporal features and high-level semantic features, we obtain scheduling fusion features. Based on the scheduling fusion features, we plan multiple resource scheduling durations, resource scheduling costs, and resource scheduling methods to obtain multiple candidate scheduling strategies corresponding to the transportation of mobile resources from the resource scheduling starting location to the resource scheduling ending location.

[0069] The attention mechanism can be used to integrate spatiotemporal traffic features with high-level semantic features. This is a fusion method that dynamically selects important information. During the feature fusion process of the heterogeneous resource graph and the traffic graph, the attention mechanism can assign different weights to resource features or traffic features based on the importance of the current task.

[0070] For each node v i , corresponding to the resource graph feature q vi (i.e., high-level semantic features) and traffic map features z vi (i.e., traffic spatiotemporal features), the fusion of the two can be achieved through attention weights α and β: i =αq vi +βz vi .

[0071] The weights α and β are calculated dynamically by the attention network according to the different tasks: Wr and Wt are the linear transformation weight matrices of the resource heterogeneity graph and traffic graph features. α and β are attention weights, indicating the relative importance of resource features and traffic features.

[0072] S150 . Determine, from a plurality of candidate scheduling strategies, a target scheduling strategy corresponding to transporting mobile resources from a resource scheduling starting position to a resource scheduling ending position based on a scheduling loss value corresponding to each candidate scheduling strategy.

[0073] In some embodiments, determining, from a plurality of candidate scheduling strategies, a target scheduling strategy corresponding to transporting a mobile resource from a resource scheduling starting location to a resource scheduling ending location based on a scheduling loss value corresponding to each candidate scheduling strategy includes:

[0074] The resource scheduling cost of each candidate scheduling strategy is converted into a scheduling loss value corresponding to each candidate scheduling strategy; based on the scheduling loss value corresponding to each candidate scheduling strategy and the preset scheduling cost, the scheduling optimization representation corresponding to each candidate scheduling strategy is determined; the candidate scheduling strategy whose scheduling optimization representation meets the preset scheduling requirements is determined to be the target scheduling strategy corresponding to the transportation of mobile resources from the resource scheduling starting position to the resource scheduling ending position.

[0075] The resource scheduling cost of each candidate scheduling strategy (i.e., the cost of each edge in the scheduling fusion feature, such as fuel cost, labor cost, etc.) can be converted into a scheduling loss value.

[0076] Scheduling optimization is expressed as the difference between the scheduling loss value corresponding to the candidate scheduling strategy and the preset scheduling cost, which can be used to represent the scheduling error of the candidate scheduling strategy.

[0077] if is the predicted scheduling cost (i.e., scheduling loss value), is the actual cost (i.e., the preset scheduling cost), then the cost-based loss function is: Assume that the true optimal scheduling method is The best scheduling method inferred from multiple graph features is Where N is the number of scheduling methods. The final loss function is L = W1L money +W2L y , W is the weight.

[0078] In this embodiment, traffic flow data between a resource scheduling start position and a resource scheduling end position is obtained, where the resource scheduling start position is the starting transportation position of the mobile resource, and the resource scheduling end position is the target transportation position of the mobile resource. From the traffic flow data between the resource scheduling start position and the resource scheduling end position, spatiotemporal traffic features between the resource scheduling start position and the resource scheduling end position are extracted. High-level semantic features are obtained in a resource heterogeneous graph, where the resource heterogeneous graph includes: multiple resource nodes and connecting edges between different resource nodes, where the resource nodes have corresponding node types, which are used to describe different types of resources, resource transportation equipment, and resource transportation methods, and where the connecting edges have corresponding relationship types, which are used to describe transportation equipment transportation resources and transportation equipment transportation methods. Based on the spatiotemporal traffic features and the high-level semantic features, multiple candidate scheduling strategies corresponding to the transportation of the mobile resource from the resource scheduling start position to the resource scheduling end position are determined, where the candidate scheduling strategies include: resource scheduling duration, resource scheduling cost, and resource scheduling method. From the multiple candidate scheduling strategies, a target scheduling strategy corresponding to the transportation of the mobile resource from the resource scheduling start position to the resource scheduling end position is determined based on the scheduling loss value corresponding to each candidate scheduling strategy. In this way, by analyzing the relationship and connection paths between different transportation nodes, the accuracy of resource scheduling can be effectively improved.

[0079] In summary, the existing rapid resource scheduling methods have problems such as complex models, ignoring heterogeneous graphs, and not considering the spatiotemporal characteristics of traffic flow graph data. In view of the spatiotemporal characteristics of traffic flow graphs, this embodiment uses a graph convolutional network to establish spatial geographic features, and an LSTM network to establish temporal traffic features, and then fuses the spatiotemporal features to represent the traffic time patterns and geographical patterns of spatial regions at different time steps; in view of resource heterogeneous graphs, resource heterogeneous graphs contain a variety of different types of nodes (resources, hubs, transportation tools, etc.) and edges (transportation relationships, scheduling relationships, etc.), which can obtain high-level semantic features of resource flow through the flow of resources in the transmission system, scheduling priorities, and complex relationships. By training the two features in combination with the time factor, the rapid resource scheduling method has rich spatiotemporal data characteristics, which greatly improves the prediction effect. Compared with previous urban traffic prediction models, this embodiment has the characteristics of simple network, small number of parameters, short training cycle, and good prediction effect.

[0080] Figure 2 This embodiment provides a structural diagram of a resource scheduling device based on a heterogeneous transportation network. The resource scheduling device based on a heterogeneous transportation network may include: a first acquisition module 210, an extraction module 220, a second acquisition module 230, a first determination module 240 and a second determination module 250.

[0081] The first acquisition module 210 is used to acquire traffic flow data between a resource scheduling start position and a resource scheduling end position, where the resource scheduling start position is the starting transportation position of the mobile resource and the resource scheduling end position is the target transportation position of the mobile resource.

[0082] The extraction module 220 is configured to extract the spatiotemporal characteristics of traffic between the resource scheduling start location and the resource scheduling end location from the traffic flow data between the resource scheduling start location and the resource scheduling end location.

[0083] The second acquisition module 230 is used to obtain high-level semantic features in the resource heterogeneous graph. The resource heterogeneous graph includes: multiple resource nodes and connection edges between different resource nodes. The resource nodes correspond to node types, and the node types are used to describe different types of resources, resource transportation equipment and resource transportation methods. The connection edges correspond to relationship types, and the relationship types are used to describe transportation equipment transportation resources and transportation equipment transportation methods.

[0084] The first determination module 240 is used to determine multiple candidate scheduling strategies corresponding to the transportation of mobile resources from the resource scheduling starting location to the resource scheduling ending location based on traffic spatiotemporal characteristics and high-level semantic characteristics. The candidate scheduling strategies include: resource scheduling duration, resource scheduling cost and resource scheduling method.

[0085] The second determination module 250 is configured to determine, from a plurality of candidate scheduling strategies, a target scheduling strategy corresponding to transporting the mobile resource from the resource scheduling starting position to the resource scheduling ending position based on the scheduling loss value corresponding to each candidate scheduling strategy.

[0086] In this embodiment, optionally, the extraction module 220 includes: a first extraction unit, a second extraction unit, and a fusion unit.

[0087] The first extraction unit is used to extract traffic space features between the resource scheduling starting position and the resource scheduling ending position from traffic flow data between the resource scheduling starting position and the resource scheduling ending position.

[0088] The second extraction unit is used to extract the traffic time feature between the resource scheduling starting position and the resource scheduling ending position from the traffic flow data between the resource scheduling starting position and the resource scheduling ending position.

[0089] The fusion unit is used to fuse the traffic space characteristics and the traffic time characteristics to obtain the traffic space-time characteristics between the resource scheduling starting position and the resource scheduling ending position.

[0090] In this embodiment, optionally, the second acquisition module 230 is specifically configured to:

[0091] A heterogeneous graph neural network is used to perform representation learning on the resource heterogeneous graph to capture the structural information and semantic information in the resource heterogeneous graph; based on the structural information and semantic information in the resource heterogeneous graph, the high-level semantic features in the resource heterogeneous graph are determined.

[0092] In this embodiment, optionally, the first determining module 240 is specifically configured to:

[0093] By integrating traffic spatiotemporal features and high-level semantic features, we obtain scheduling fusion features. Based on the scheduling fusion features, we plan multiple resource scheduling durations, resource scheduling costs, and resource scheduling methods to obtain multiple candidate scheduling strategies corresponding to the transportation of mobile resources from the resource scheduling starting location to the resource scheduling ending location.

[0094] In this embodiment, optionally, the second determining module 250 is specifically configured to:

[0095] The resource scheduling cost of each candidate scheduling strategy is converted into a scheduling loss value corresponding to each candidate scheduling strategy; based on the scheduling loss value corresponding to each candidate scheduling strategy and the preset scheduling cost, the scheduling optimization representation corresponding to each candidate scheduling strategy is determined; the candidate scheduling strategy whose scheduling optimization representation meets the preset scheduling requirements is determined to be the target scheduling strategy corresponding to the transportation of mobile resources from the resource scheduling starting position to the resource scheduling ending position.

[0096] In this embodiment, optionally, the first extraction unit is specifically configured to:

[0097] A graph convolutional network is used to extract the spatial representation information corresponding to the traffic flow data between the resource scheduling start position and the resource scheduling end position; based on the spatial representation information, the traffic spatial characteristics between the resource scheduling start position and the resource scheduling end position are determined.

[0098] In this embodiment, optionally, the second extraction unit is specifically configured to:

[0099] A long short-term memory network is used to extract the time representation information corresponding to the traffic flow data between the resource scheduling starting position and the resource scheduling ending position; based on the time representation information, the traffic time characteristics between the resource scheduling starting position and the resource scheduling ending position are determined.

[0100] The resource scheduling device based on the heterogeneous transportation network provided by the present disclosure can execute the above method embodiments. Its specific implementation principles and technical effects can be found in the above method embodiments, and the present disclosure will not repeat them here.

[0101] The present application also provides a computer device. Figure 3 , Figure 3 This is a basic structural block diagram of the computer device in this embodiment.

[0102] The computer device includes a memory 310 and a processor 320 that are interconnected and communicate with each other via a system bus. It should be noted that the figure only shows a computer device with a memory 310 and a processor 320, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to a microprocessor, an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a digital signal processor (DSP), an embedded device, etc.

[0103] Computer devices can be desktop computers, laptops, PDAs, cloud servers, etc. Computer devices can interact with users through keyboards, mice, remote controls, touchpads, or voice-activated devices.

[0104] The memory 310 includes at least one type of readable storage medium, including non-volatile memory or volatile memory, such as flash memory, hard disk, multimedia card, card-type memory (such as SD or DX memory), random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic storage, magnetic disk, optical disk, etc. The RAM may include static RAM or dynamic RAM. In some embodiments, the memory 310 may be an internal storage unit of a computer device, such as a hard disk or memory of the computer device. In other embodiments, the memory 310 may also be an external storage device of the computer device, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, or a Flash Card equipped on the computer device. Of course, the memory 310 may also include both the internal storage unit of the computer device and its external storage device. In this embodiment, the memory 310 is generally used to store the operating system and various application software installed on the computer device, such as the program code of the above-mentioned method. In addition, the memory 310 may also be used to temporarily store various types of data that have been output or are about to be output.

[0105] The processor 320 is generally used to perform the overall operation of the computer device. In this embodiment, the memory 310 is used to store program code or instructions, which include computer operating instructions. The processor 320 is used to execute the program code or instructions stored in the memory 310 or process data, such as the program code for running the above method.

[0106] In this document, a bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. This bus system can be divided into address buses, data buses, and control buses. For ease of illustration, the figure uses only one thick line, but this does not mean that there is only one bus or only one type of bus.

[0107] Another embodiment of the present application further provides a computer-readable medium, which may be a computer-readable signal medium or a computer-readable medium. A processor in a computer reads the computer-readable program code stored in the computer-readable medium, enabling the processor to execute the functional actions specified in each step or combination of steps in the above method, and to generate a device that implements the functional actions specified in each block or combination of blocks in the block diagram.

[0108] Computer-readable media include but are not limited to electronic, magnetic, optical, electromagnetic, infrared memory or semiconductor systems, devices or apparatuses, or any appropriate combination of the foregoing, the memory is used to store program codes or instructions, the program codes include computer operating instructions, and the processor is used to execute the program codes or instructions of the above-mentioned methods stored in the memory.

[0109] For the definitions of memory and processor, please refer to the description of the aforementioned computer device embodiment and will not be repeated here.

[0110] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of modules or units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0111] Each functional unit or module in each embodiment of the present application may be integrated into a processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The above-mentioned integrated units may be implemented in the form of hardware or software functional units.

[0112] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) or a processor (processor) to execute all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program code.

[0113] In the claims, any reference signs placed between brackets shall not be construed as limiting the claims. The word "comprising" described in the present application does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present application can be implemented with the aid of hardware comprising several different elements and with the aid of a suitably programmed computer. In a unit claim that lists several means, several units of these means may be embodied by the same hardware item. The use of first, second, and third etc. does not indicate any order and these words may be interpreted as names. The steps in the above embodiments should not be understood as limiting the order of execution unless otherwise specified.

[0114] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A resource scheduling method based on heterogeneous transportation networks, characterized in that: include: Acquire traffic flow data between a resource scheduling starting position and a resource scheduling ending position, wherein the resource scheduling starting position is a starting transportation position of the mobile resource and the resource scheduling ending position is a target transportation position of the mobile resource; Extracting the spatiotemporal characteristics of traffic between the resource scheduling starting position and the resource scheduling ending position from the traffic flow data between the resource scheduling starting position and the resource scheduling ending position; Obtaining high-level semantic features in a resource heterogeneous graph, wherein the resource heterogeneous graph includes: a plurality of resource nodes and connection edges between different resource nodes, wherein the resource nodes have corresponding node types, wherein the node types are used to describe different types of resources, resource transportation equipment, and resource transportation methods, and wherein the connection edges have corresponding relationship types, wherein the relationship types are used to describe transportation equipment transportation resources and transportation equipment transportation methods; Based on the traffic spatiotemporal characteristics and the high-level semantic characteristics, determining a plurality of candidate scheduling strategies corresponding to the transportation of the mobile resource from the resource scheduling starting location to the resource scheduling ending location, the candidate scheduling strategies including: resource scheduling duration, resource scheduling cost, and resource scheduling mode; From the plurality of candidate scheduling strategies, based on the scheduling loss value corresponding to each candidate scheduling strategy, a target scheduling strategy corresponding to the transportation of the mobile resource from the resource scheduling starting position to the resource scheduling ending position is determined.

2. The method according to claim 1, characterized in that Extracting the spatiotemporal characteristics of traffic between the resource scheduling starting position and the resource scheduling ending position from the traffic flow data between the resource scheduling starting position and the resource scheduling ending position includes: Extracting traffic space characteristics between the resource scheduling starting position and the resource scheduling ending position from traffic flow data between the resource scheduling starting position and the resource scheduling ending position; Extracting traffic time characteristics between the resource scheduling starting position and the resource scheduling ending position from traffic flow data between the resource scheduling starting position and the resource scheduling ending position; The traffic space characteristics and the traffic time characteristics are integrated to obtain the traffic space-time characteristics between the resource scheduling starting position and the resource scheduling ending position.

3. The method according to claim 1, characterized in that The step of obtaining high-level semantic features in the resource heterogeneous graph includes: Using a heterogeneous graph neural network to perform representation learning on the resource heterogeneous graph, thereby capturing structural information and semantic information in the resource heterogeneous graph; Based on the structural information and semantic information in the resource heterogeneous graph, high-level semantic features in the resource heterogeneous graph are determined.

4. The method according to claim 1, wherein The determining, based on the traffic spatiotemporal features and the high-level semantic features, a plurality of candidate scheduling strategies corresponding to transporting the mobile resource from the resource scheduling starting location to the resource scheduling ending location includes: fusing the traffic spatiotemporal features and the high-level semantic features to obtain a scheduling fusion feature; Based on the scheduling fusion feature, multiple resource scheduling durations, resource scheduling costs and resource scheduling methods are planned to obtain multiple candidate scheduling strategies corresponding to the transportation of the mobile resource from the resource scheduling starting position to the resource scheduling ending position.

5. The method according to claim 1, wherein The determining, from the plurality of candidate scheduling strategies, a target scheduling strategy corresponding to transporting the mobile resource from the resource scheduling starting position to the resource scheduling ending position based on a scheduling loss value corresponding to each candidate scheduling strategy, includes: Converting the resource scheduling cost of each candidate scheduling strategy into a scheduling loss value corresponding to each candidate scheduling strategy; Determining a scheduling optimization representation corresponding to each candidate scheduling strategy based on a scheduling loss value and a preset scheduling cost corresponding to each candidate scheduling strategy; Determining the scheduling optimization indicates that the candidate scheduling strategy that meets the preset scheduling requirements is the target scheduling strategy corresponding to the transportation of the mobile resource from the resource scheduling starting position to the resource scheduling ending position.

6. The method according to claim 2, characterized in that Extracting the traffic space characteristics between the resource scheduling starting position and the resource scheduling ending position from the traffic flow data between the resource scheduling starting position and the resource scheduling ending position includes: Using a graph convolutional network, extracting spatial representation information corresponding to traffic flow data between the resource scheduling start location and the resource scheduling end location; Based on the spatial representation information, the traffic spatial characteristics between the resource scheduling starting position and the resource scheduling ending position are determined.

7. The method according to claim 2, characterized in that The extracting, from the traffic flow data between the resource scheduling starting position and the resource scheduling ending position, the traffic time feature between the resource scheduling starting position and the resource scheduling ending position comprises: Using a long short-term memory network, extracting time representation information corresponding to traffic flow data between the resource scheduling start location and the resource scheduling end location; Based on the time representation information, a traffic time characteristic between the resource scheduling start position and the resource scheduling end position is determined.

8. A resource scheduling device based on a heterogeneous transportation network, characterized in that: include: A first acquisition module is configured to acquire traffic flow data between a resource scheduling start position and a resource scheduling end position, wherein the resource scheduling start position is a starting transportation position of a mobile resource and the resource scheduling end position is a target transportation position of the mobile resource; an extraction module, configured to extract, from the traffic flow data between the resource scheduling starting position and the resource scheduling ending position, the spatiotemporal characteristics of traffic between the resource scheduling starting position and the resource scheduling ending position; A second acquisition module is configured to acquire high-level semantic features in a resource heterogeneous graph, wherein the resource heterogeneous graph includes: a plurality of resource nodes and connection edges between different resource nodes, wherein the resource nodes have corresponding node types, wherein the node types are used to describe different types of resources, resource transportation equipment, and resource transportation methods, and wherein the connection edges have corresponding relationship types, wherein the relationship types are used to describe transportation equipment transportation resources and transportation equipment transportation methods; A first determination module is configured to determine, based on the traffic spatiotemporal characteristics and the high-level semantic characteristics, a plurality of candidate scheduling strategies corresponding to the transportation of the mobile resource from the resource scheduling starting location to the resource scheduling ending location, the candidate scheduling strategies including: resource scheduling duration, resource scheduling cost, and resource scheduling mode; The second determination module is used to determine the target scheduling strategy corresponding to the transportation of the mobile resource from the resource scheduling starting position to the resource scheduling ending position from the multiple candidate scheduling strategies based on the scheduling loss value corresponding to each candidate scheduling strategy.

9. A computer device, characterized in that: The system comprises a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the resource scheduling method based on a heterogeneous transportation network as claimed in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the resource scheduling method based on a heterogeneous transportation network as claimed in any one of claims 1 to 7 is implemented.

Images