Intelligent allocation method and system for operation resources on sanitation cloud platform

Abstract

This invention provides a method and system for intelligently allocating operational resources on a cloud sanitation platform. The method comprises the following steps: a data input step for collecting raw data; a data cleaning step for cleaning the collected raw data to form a data blueprint; an event processing step for pre-planning tasks based on the data blueprint to generate partial processing events; an emergency event reporting step based on the data blueprint; and a resource allocation step for real-time resource allocation based on reported emergencies. This invention utilizes cutting-edge computer technologies such as deep learning, model feature extraction, road prediction, and trajectory fitting to implement a standardized process for cloud sanitation resource allocation, avoiding the subjectivity and limitations of manual allocation.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent sanitation, and in particular to a method and system for intelligent allocation of operating resources on a sanitation cloud platform. Background Art

[0002] With the rise of cloud services, the cloudification of traditional industries has gradually become the mainstream of industrial transformation. Currently, in the sanitation sector, cloud-based services are continuously advancing. Leveraging the real-time and interconnected advantages of cloud platforms, traditional sanitation work has seen significant improvements in labor costs, work efficiency, and work quality.

[0003] In the daily operations of cloud platforms, the allocation of sanitation resources is a long-standing and difficult problem. The main contradiction in sanitation resource allocation lies in how to match limited sanitation resources with pending events. Limited sanitation resources refer to the fixed total number of sanitation vehicles, fixed locations of maintenance points, and a fixed number of maintenance personnel. Pending events are mainly divided into two aspects: one is advance task planning, such as implementing a specific cleaning mode for a road section or intersection when rainfall reaches a certain threshold on rainy days; the other is unpredictable accidents, such as a car accident at an intersection, where the road is blocked by debris and garbage, affecting traffic safety. The current sanitation cloud platform's resource classification system aims to meet these requirements.

[0004] The Chinese invention patent document with publication number CN115438971A discloses a sanitation operation management method and system thereof, wherein the sanitation operation management method includes the following steps: S110: obtaining the trajectory data of the operation vehicle in a preset time period, executing S120; S120: querying all operation types in the preset time period, taking all operation types as a set, selecting any operation type in the set, and executing S130; S130: reading the road basic data of all roads, obtaining the statistical information of each road based on the road basic data, trajectory data and operation status data, and executing S140; S140: normalizing the statistical information of all roads to obtain an assembled structured object, adding the assembled structured object to the data set to be stored, and executing S150; S150: selecting another operation type from all operation types, executing S130, until all operation types complete the cyclic calculation, and storing and displaying the data set to be stored as set data. The Chinese invention patent document with publication number CN115358589A discloses a method, device, equipment and storage medium for scheduling autonomous sanitation vehicles. The method includes: obtaining several task routes, vehicle data of the autonomous sanitation vehicles to be scheduled, and resource consumption of the autonomous sanitation vehicles to be scheduled in different operating modes; calculating a scheduling plan with the shortest operation time based on the task routes, resource consumption of the autonomous sanitation vehicles to be scheduled in different operating modes and vehicle data; obtaining task allocation results according to the scheduling plan, and sending the task allocation results to the corresponding on-board server through the cloud server, so that the autonomous sanitation vehicles to be scheduled perform sanitation operations according to the task allocation results.

[0005] Regarding the above-mentioned related technologies, the inventor believes that in solving the problem of low efficiency in the management and scheduling of sanitation vehicles by staff, only the resource consumption of different vehicle operation modes and the calculation of the scheduling plan with the shortest operation time are simply considered, ignoring the impact of emergencies, operating road conditions and matching the optimal resource allocation method that meets user requirements on the resource allocation and scheduling of the intelligent platform. Summary of the Invention

[0006] In view of the defects in the prior art, the purpose of the present invention is to provide a method and system for intelligent allocation of operating resources of a sanitation cloud platform.

[0007] According to the present invention, a method for intelligently allocating operating resources on a sanitation cloud platform is provided, comprising the following steps:

[0008] Data entry steps: Collect raw data;

[0009] Data cleaning step: clean the collected raw data to form a data blueprint;

[0010] Incident handling steps: Plan tasks in advance according to the data blueprint to form some processing events; report emergencies according to the data blueprint;

[0011] Resource allocation step: Real-time resource allocation for reported emergencies.

[0012] Preferably, in the resource allocation step, the content of the event to be processed is aggregated by starting and ending points, operation event segments, and operation modes, the event content is marked, the resource library is searched and sorted, and the overall layout of the current resources is obtained by analyzing the resources that have not been used, the resources in operation, and the operation points;

[0013] According to the overall layout of current resources, different matching operators and factors are used to fit and calculate the actual operation road.

[0014] Preferably, in the resource allocation step, integration of events to be processed is performed;

[0015] Aggregate and correlate events to be processed in multiple dimensions;

[0016] Set up pre-known tasks for planned events in the processing events, and configure the scope and method of operation in advance by the system administrator;

[0017] For real-time events in the processing events, pre-analyze the events and then issue instructions;

[0018] After real-time events and planned events enter resource allocation, they are labeled from multiple dimensions to form a feature map of the events.

[0019] Preferably, in the resource allocation step, the running resources and the unused resources in the resource library are retrieved; the dynamic and static resources are evaluated in real time to obtain the overall operating capacity of the current system;

[0020] Operating resources are objects that are currently performing sanitation work; non-operating resources are stock objects on standby, and the subjects of stock objects include but are not limited to vehicles, operation and maintenance points, and checkpoints.

[0021] Preferably, in the resource allocation step, operation trajectory fitting is performed to comprehensively match the current total amount of resources with the events to be processed;

[0022] Abstract the reference factors of resource allocation into factors; combine factors, ops, and values ​​into semantic expressions. The combination of multiple semantic expressions forms a factor set. When resources are actually allocated, the allocation tendency is obtained by reading the content of the factor set.

[0023] Taking into account the current status of all resources, the traffic planning prediction model is used to predict the actual commuting conditions of each feasible topological route within a specified time period, and relevant convolution operations are performed in the time dimension to obtain the model solution.

[0024] According to the present invention, an intelligent resource allocation system for sanitation cloud platform operations is provided, which includes the following modules:

[0025] Data input module: collect raw data;

[0026] Data cleaning module: cleans the collected raw data to form a data blueprint;

[0027] Event processing module: carry out advance task planning according to the data blueprint to form some processing events; report emergencies according to the data blueprint;

[0028] Resource allocation module: real-time resource allocation for reported emergencies.

[0029] Preferably, in the resource allocation module, the content of the event to be processed is aggregated by starting and ending points, operation event segments, and operation modes, the event content is marked, the resource library is searched and sorted, and the overall layout of the current resources is obtained by analyzing the resources that have not been used, the resources in operation, and the operation points;

[0030] According to the overall layout of current resources, different matching operators and factors are used to fit and calculate the actual operation road.

[0031] Preferably, in the resource allocation module, integration of pending events is performed;

[0032] Aggregate and correlate events to be processed in multiple dimensions;

[0033] Set up pre-known tasks for planned events in the processing events, and configure the scope and method of operation in advance by the system administrator;

[0034] For real-time events in the processing events, pre-analyze the events and then issue instructions;

[0035] After real-time events and planned events enter resource allocation, they are labeled from multiple dimensions to form a feature map of the events.

[0036] Preferably, in the resource allocation module, the running resources and the unused resources in the resource library are retrieved; the dynamic and static resources are evaluated in real time to obtain the overall operating capacity of the current system;

[0037] Operating resources are objects that are currently performing sanitation work; non-operating resources are stock objects on standby, and the subjects of stock objects include but are not limited to vehicles, operation and maintenance points, and checkpoints.

[0038] Preferably, in the resource allocation module, operation trajectory fitting is performed to coordinately match the current total amount of resources with the events to be processed;

[0039] Abstract the reference factors of resource allocation into factors; combine factors, ops, and values ​​into semantic expressions. The combination of multiple semantic expressions forms a factor set. When resources are actually allocated, the allocation tendency is obtained by reading the content of the factor set.

[0040] Taking into account the current status of all resources, the traffic planning prediction model is used to predict the actual commuting conditions of each feasible topological route within a specified time period, and relevant convolution operations are performed in the time dimension to obtain the model solution.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] 1. This invention uses cutting-edge computer technologies such as deep learning, model feature extraction, road prediction, and trajectory fitting to implement a standard process for cloud sanitation resource allocation, avoiding the subjectivity and limitations of manual allocation;

[0043] 2. In this invention, managers only need to set relevant factors and thresholds. The system will intelligently allocate time for pending events based on the factors. The three basic integrated capabilities currently include time optimization, which uses road busyness prediction and route topology to calculate the running time of each possible route; distance optimization, which uses trajectory topology and local optimal algorithms to obtain the shortest route; and resource optimization, which uses the CNN network model evolution on the resource pool and factor constraints to obtain the optimal resource implementation solution.

[0044] 3. In addition to the above basic capabilities, the invention also provides good scalability. Users can achieve more diversified resource allocation scenarios by customizing factors and introducing new capabilities through interfaces. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments with reference to the following drawings:

[0046] Figure 1 This is the flow chart of the intelligent allocation system for operating resources on the sanitation cloud platform;

[0047] Figure 2 It is a real-time event processing flow chart;

[0048] Figure 3Implementation diagram for job fitting;

[0049] Figure 4 Analyze the diagram for detailed processing of events;

[0050] Figure 5 The following diagram shows the distribution effect. DETAILED DESCRIPTION

[0051] The present invention will be described in detail below with reference to specific embodiments. The following examples will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those skilled in the art, several changes and improvements can be made without departing from the scope of the present invention. These all fall within the scope of protection of the present invention.

[0052] The embodiment of the present invention discloses a method and system for intelligent allocation of sanitation cloud platform operation resources, which is a solution for sanitation operation resource allocation, such as Figure 1 As shown, the specific steps include:

[0053] Data entry step: Collect raw data.

[0054] Data cleaning step: Clean the collected raw data to form a data blueprint.

[0055] Event handling steps: Plan tasks in advance according to the data blueprint to form some processing events; report emergencies according to the data blueprint.

[0056] Starting with raw data collection and cleaning to create a usable data blueprint, this data blueprint serves as the basis for advance task planning, thereby forming part of the event handling process. For emergencies, real-time resource allocation is implemented through event reporting. Raw data includes map acquisition data, point cloud mapping data, and real-world sampling data. Data input, data cleaning, and operation planning are all technical solutions in this field, and our company has patents to further explore and protect them. Examples of emergencies include car accidents, road obstructions, water leaks at water supply stations, and other factors that affect operation planning.

[0057] Resource allocation step: Real-time resource allocation for reported emergencies.

[0058] During the actual resource allocation process, the event content is aggregated by starting and ending points, operation event segments, and operation methods. The event content itself is labeled, and the resource library is searched and organized. The overall layout of the current resources is analyzed from the perspectives of unused resources, running resources, and operation points. Finally, different matching operators and factors are used to fit and calculate the actual operation path. For example, when adopting the optimal distance solution, adjacent point matching and road backtracking are required for search and processing. The actual operation path fitting and calculation involves contents such as operation mode, operation time, path planning, energy supply, etc. The present invention performs different calculations based on the search of the resource library and the configuration of factor sets for the events to be processed, avoiding the unscientific nature of manual planning.

[0059] Integration of pending events: Integration of pending events is the aggregation and correlation of events in multiple dimensions.

[0060] To-be-processed events consist of planned events and real-time events. Planned events are tasks that are known in advance and are generally configured in advance by the system administrator, with a clear scope and method of operation. Real-time events, on the other hand, refer to random events that occur on the road in daily life, generally related to car accidents and natural weather. These events themselves do not have a clear operation plan, and the system needs to pre-analyze the events before issuing clear instructions.

[0061] Therefore, there is an event parsing engine between emergencies and resource allocation to achieve unified resource allocation input, such as Figure 2 As shown in the figure, in a job scheduling system, the event parsing module specifically includes event classification, feature recognition, similarity analysis, and resource matching. Planned events are pre-determined instructions that can directly trigger resource allocation. However, emergencies lack specific instructions, so a preliminary event parsing step is required to convert uncertain events into instructions that can be used for resource allocation.

[0062] After real-time events and planned events enter resource allocation, they will be labeled based on dimensions such as start and end points, operation time period, operation vehicle type, and operation method to form a feature map of the events.

[0063] Resource Library Search: Resources in the resource library include both active and inactive resources. Active resources refer to objects currently performing sanitation work, while inactive resources refer to objects on standby. These objects include, but are not limited to, vehicles, maintenance points, and checkpoints. Resource library search evaluates these dynamic and static resources in real time to determine the overall operational capacity of the current system.

[0064] Job trajectory fitting: Job trajectory fitting is the process of coordinating and matching the current total amount of resources with the events to be processed. The reference factors of resource allocation are abstracted into the concept of factors. Factors, op (options) and value (option values) constitute a semantic expression. The combination of multiple semantic expressions forms a factor set. When resources are actually allocated, the content in the factor set is read to obtain the tendency of this allocation. For example, if the option is biased towards a shorter time, the location information of all current resources and the location information of the events to be processed will be calculated in the resource library for the point center range, and the relevant value configured in the factor will be combined to achieve filtering within the specified range; then the actual commuting conditions of each feasible topological route within the specified time period are predicted through the road condition planning prediction model, and related convolution operations are performed on the time dimension to obtain the optimal solution of the model. Of course, the current status of all resources is also one of the considerations when planning the route. For example, the nearest vehicle A can immediately go to the emergency location L for operation processing, but the oil and electricity resources of vehicle A can only support one-way operations and cannot subsequently reach the nearest supply area. In this case, the allocation plan will not be issued. Figure 3 shown.

[0065] The resource optimization solution in this invention is based on the current state of existing resources, including all operating vehicles, their operating methods, vehicle stops, vehicle maintenance points, and vehicle resupply points. When an unexpected incident occurs near a routine operating vehicle, dispatching additional vehicles to handle it would be a significant waste of time and resources. Current operating vehicles can dynamically provide nearby support without violating operational priorities. This approach significantly reduces overall sanitation costs and fully leverages the cloud platform's comprehensive monitoring capabilities.

[0066] Figure 4 This is a specific case (processing event). In the figure, the pending job event is marked. The intelligent resource allocation system is used for solution analysis (using the default factor set configuration):

[0067] 1. When time is optimal, a work vehicle is immediately dispatched from the operation and maintenance command center along the main road (the road with the shortest predicted commuting time) to the work site at a constant speed and return after completing the work.

[0068] 2. Optimal distance: Through local resource matching, the nearest vehicle 39197 is found. After it completes its own operation, it arrives at the waiting location to complete the current operation.

[0069] 3. Optimal resources: Through resource matching, it was discovered that vehicle 36387 was heading to the operation and maintenance point to refill water. Its original route was changed, and it was first required to complete the temporary operation at the work site before going to the water filling point in Area 1 to complete the water filling work.

[0070] The above is a specific allocation plan generated under the default factor set. However, in actual application scenarios, there is rarely an absolutely single optimal solution. The solution is generally a compromise in multiple dimensions. In this case, users only need to add corresponding custom factors and configure the combination to implement a personalized solution. For example, the scheduling distance is controlled within 3km, the completion time cannot exceed 30 minutes, and the resource consumption does not exceed two autonomous vehicles.

[0071] The intelligent resource allocation in this invention constructs the existing resource pool graph structure, combines the content of the event to be processed and the custom factor set, and uses the machine capability to dynamically, quickly and in real time calculate the allocation plan that meets the requirements, such as Figure 4 shown.

[0072] Application display: The present invention shields the user from the complex matching process at the application level. The user can obtain the corresponding recommended allocation method by clicking on the relevant application interface, such as Figure 5 shown.

[0073] (1) Structuring of pending events: Both emergencies and planned events are analyzed by the parsing engine to form unified structured data, such as time period, event level, pending operation mode, number of operations, whether secondary inspection is required, etc.

[0074] (2) The concept of resource allocation factors: the user's allocation intention is concretized into actual factor objects, and flexible allocation strategies are achieved through customized operations on factors.

[0075] (3) In resource allocation, recommendations based on the road distance dimension are realized based on the matching of adjacent points calculated by path calculation and the road backtracking capability; recommendations based on the operation time dimension are realized based on the road condition prediction; and recommendations based on the resource consumption dimension are realized based on the resource selection CNN model.

[0076] The intelligent sanitation cloud resource allocation system introduced in this invention can coordinate the overall resource situation of the current system through customized factors and confidence levels, matching the optimal resource allocation method that meets user requirements, such as optimal distance, optimal resources, and optimal time. This system reduces the occurrence of irrational manual resource allocation and, through scientific model prediction and planning, greatly improves the efficiency of sanitation cloud resource allocation and reduces sanitation costs. The present invention is applied to the allocation of operational resources in sanitation cloud platform projects. Through analysis of pending events and resource library retrieval, it forms a fitting of actual operational trajectories. In the actual operational path fitting, by customizing different factors and options, a resource allocation solution that best meets user expectations is achieved.

[0077] Traditional sanitation resource allocation is mainly manual real-time scheduling, and the model adopted is as follows: when an event to be processed occurs, the operation and maintenance personnel of the cloud platform receive the corresponding notification at the management center, and then divide the resources according to the relevant attributes of the current event. For example, the following allocation method may be adopted for A1 car to go to R1 road to perform T1 type tasks, but in fact there are more than 100 allocation methods that can achieve the current time processing, and the method adopted by the operator is generally based on experience and is highly subjective. Obviously, such experience allocation is often not optimal, and there are cases where different operators have different processing methods for the same problem. On this basis, the present invention provides a set of standard solutions for sanitation resource allocation through a two-way evaluation of existing resources and events, combined with influencing factors, road prediction, route fitting, and feature fusion capabilities.

[0078] The present invention also provides an intelligent allocation system for operating resources of a sanitation cloud platform. The intelligent allocation system for operating resources of a sanitation cloud platform can be implemented by executing the process steps of the intelligent allocation method for operating resources of a sanitation cloud platform. That is, those skilled in the art can understand the intelligent allocation method for operating resources of a sanitation cloud platform as an optimal implementation method of the intelligent allocation system for operating resources of a sanitation cloud platform.

[0079] The system includes the following modules: Data input module: collects raw data.

[0080] Data cleaning module: cleans the collected raw data to form a data blueprint.

[0081] Event processing module: Plan tasks in advance according to the data blueprint to form some processing events; report emergencies according to the data blueprint.

[0082] Resource Allocation Module: This module performs real-time resource allocation for reported emergencies. It aggregates the content of pending events by starting and ending points, operation event segments, and operation methods, tags the event content, searches and organizes the resource library, and analyzes unused resources, active resources, and operation points to determine the overall current resource layout. Based on this overall resource layout, it uses different matching operators and factors to fit and calculate the actual operation path.

[0083] Integrate pending events; aggregate and associate pending events in multiple dimensions; set pre-known agents for planned events in the pending events, and configure the tasks, operation scope and operation method in advance by the system administrator; for real-time events in the pending events, pre-parse the events and then issue instructions; after real-time events and planned events enter resource allocation, mark the events from multiple dimensions to form a feature map of the events.

[0084] Retrieve the running resources and non-operating resources in the resource library; conduct real-time evaluation of dynamic and static resources to obtain the overall operating capacity of the current system; running resources are objects that are currently performing sanitation work; non-operating resources are stock objects on standby, and the subjects of stock objects include but are not limited to vehicles, operation and maintenance points, and checkpoints.

[0085] Perform job trajectory fitting to coordinately match the current total amount of resources with pending events; abstract the reference factors for resource allocation into factors; construct semantic expressions of factors, ops, and values; and combine multiple semantic expressions to form a factor set. When actually allocating resources, the allocation tendency is obtained by reading the content of the factor set; taking into account the current status of all resources, the actual commuting conditions for each feasible topological route within a specified time period are predicted through the road condition planning prediction model, and relevant convolution operations are performed on the time dimension to obtain the model solution.

[0086] Those skilled in the art will appreciate that, in addition to implementing the system and its various devices, modules, and units provided by the present invention in purely computer-readable program code, it is entirely possible to implement the same functions of the system and its various devices, modules, and units provided by the present invention in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, the system and its various devices, modules, and units provided by the present invention can be considered a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; the devices, modules, and units for implementing various functions can also be considered as both software modules implementing the method and structures within the hardware component.

[0087] The above describes specific embodiments of the present invention. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art may make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. The embodiments of this application and the features in the embodiments may be combined with each other in any manner unless there is a conflict.

Claims

1. A method for intelligent allocation of operating resources on a sanitation cloud platform, characterized in that: The steps include: Data entry steps: Collect raw data; Data cleaning step: clean the collected raw data to form a data blueprint; Event processing steps: advance task planning based on the data blueprint to form partial processing events; Report emergencies based on the data blueprint; Resource allocation step: Real-time resource allocation for reported emergencies; In the resource allocation step, the content of the event to be processed is aggregated by starting and ending points, operation event segments, and operation methods. The event content is marked, and the resource library is searched and sorted. The current overall resource layout is obtained by analyzing the resources that have not been used, the resources that are in operation, and the operation points. According to the overall layout of current resources, different matching operators and factors are used to fit and calculate the actual operation road; In the resource allocation step, integrating events to be processed; Aggregate and correlate events to be processed in multiple dimensions; Set up pre-known tasks for planned events in the processing events, and configure the scope and method of operation in advance by the system administrator; For real-time events in the processing events, pre-analyze the events and then issue instructions; After real-time events and planned events enter resource allocation, they are labeled from multiple dimensions to form a feature map of the events. In the resource allocation step, the running resources and inactive resources in the resource library are retrieved; dynamic and static resources are evaluated in real time to obtain the overall operating capacity of the current system; Active resources are objects currently performing sanitation work; inactive resources are inventory objects on standby, including but not limited to vehicles, operation and maintenance points, and checkpoints. In the resource allocation step, job trajectory fitting is performed to coordinate and match the current total amount of resources with the events to be processed; Abstract the reference factors of resource allocation into factors; combine factors, ops, and values ​​into semantic expressions. The combination of multiple semantic expressions forms a factor set. When resources are actually allocated, the allocation tendency is obtained by reading the content of the factor set. Taking into account the current status of all resources, the traffic planning prediction model predicts the actual commuting conditions for each feasible topological route within a specified time period, and performs relevant convolution operations on the time dimension to obtain the model solution. By predicting road busyness, the route topology calculates the running time of each possible route; Through trajectory topology, the local optimal algorithm obtains the shortest route; Through the model evolution of the CNN network on the resource pool, the optimal resource implementation solution is obtained in conjunction with factor constraints.

2. An intelligent allocation system for sanitation cloud platform operation resources, characterized in that: Includes the following modules: Data input module: collect raw data; Data cleaning module: cleans the collected raw data to form a data blueprint; Event processing module: performs advance task planning based on the data blueprint to form some processing events; Report emergencies based on the data blueprint; Resource allocation module: real-time resource allocation for reported emergencies; In the resource allocation module, the content of the event to be processed is aggregated by starting and ending points, operation event segments, and operation methods. The event content is marked, and the resource library is searched and sorted. The current overall resource layout is obtained by analyzing the resources that have not been used, the resources that are in operation, and the operation points. According to the overall layout of current resources, different matching operators and factors are used to fit and calculate the actual operation road; In the resource allocation module, integrating events to be processed; Aggregate and correlate events to be processed in multiple dimensions; Set up pre-known tasks for planned events in the processing events, and configure the scope and method of operation in advance by the system administrator; For real-time events in the processing events, pre-analyze the events and then issue instructions; After real-time events and planned events enter resource allocation, they are labeled from multiple dimensions to form a feature map of the events. In the resource allocation module, the running resources and inactive resources in the resource library are retrieved; dynamic and static resources are evaluated in real time to obtain the overall operating capacity of the current system; Active resources are objects currently performing sanitation work; inactive resources are inventory objects on standby, including but not limited to vehicles, operation and maintenance points, and checkpoints. In the resource allocation module, job trajectory fitting is performed to coordinate and match the current total amount of resources with the events to be processed; Abstract the reference factors of resource allocation into factors; combine factors, ops, and values ​​into semantic expressions. The combination of multiple semantic expressions forms a factor set. When resources are actually allocated, the allocation tendency is obtained by reading the content of the factor set. Taking into account the current status of all resources, the traffic planning prediction model predicts the actual commuting conditions for each feasible topological route within a specified time period, and performs relevant convolution operations on the time dimension to obtain the model solution. By predicting road busyness, the route topology calculates the running time of each possible route; Through trajectory topology, the local optimal algorithm obtains the shortest route; Through the model evolution of the CNN network on the resource pool, the optimal resource implementation solution is obtained in conjunction with factor constraints.

Images