Intelligent city parking resource prediction and optimization decision-making platform based on big data

Abstract

The invention relates to the technical field of parking resource management, in particular to an intelligent city parking resource prediction and optimization decision-making platform based on big data, and the platform comprises a parking space-time modeling module, a demand potential recognition module, a resource reuse prediction module, a parking space allocation decision-making module, and a trend feedback optimization module. According to the method, the vehicle space-time behaviors are extracted, the continuous parking points are judged, the recognition precision of microscopic parking distribution is improved, the parking density sequence is constructed, density grade division is introduced, differentiated description and dynamic grading of regional parking requirements are achieved, the maximum allowable parking time is compared with the actual occupation time, and the parking density is improved. The method comprises the following steps: identifying reusable time period parking space resources, excavating parking space reuse potential, calibrating and matching parking spaces and requests based on a time length difference value and a time tolerance, enhancing scheduling accuracy and response timeliness, introducing a supply-demand proportion pushing priority strategy, enabling the parking resources to incline to a high-demand area, and integrally enhancing the fineness and timeliness of urban parking resource configuration.

Description

Technical Field

[0001] The present invention relates to the technical field of parking resource management, and in particular to a smart city parking resource prediction and optimization decision-making platform based on big data. Background Art

[0002] The field of parking resource management technology encompasses the research and application of resource allocation and scheduling methods related to parking in cities. Its core focus is the dynamic monitoring, rational allocation, and scheduling optimization of limited urban parking resources to alleviate urban traffic pressure and improve resource utilization. Key technical areas within this field include urban traffic management, dynamic resource allocation, information collection and processing, and user behavior analysis. These encompass a variety of data-driven and decision-support mechanisms, serving the overall intelligent management needs of urban transportation.

[0003] Among them, the smart city parking resource prediction and optimization decision-making platform based on big data refers to the use of big data analysis technology to collect, integrate and model multi-source parking-related data in the city, so as to predict parking demand in the future and generate corresponding resource optimization allocation plans based on this. The main technical matters include: building a prediction model based on parking history records, geographic location, time dimension and traffic flow; formulating resource allocation strategies based on prediction results; implementing resource management based on vehicle travel characteristics and regional supply and demand relationships. The generally used data collection methods include road monitoring system data, parking lot entry and exit records, mobile device location data and traffic control center information integration, and feature extraction and law modeling through data fusion methods, and finally completing the prediction and optimization plan formulation process.

[0004] Existing urban parking management systems suffer from a crude problem of identifying parking behavior zones. Parking analysis often relies on macro-aggregate statistical methods, which struggle to capture micro-level parking behaviors characterized by high continuity and high density. This leads to the neglect of high-frequency areas in policy formulation, hampering the agility of management responses. Parking demand intensity is often identified using a single numerical metric, lacking a stratified approach to different density levels. Resource regulation strategies cannot accurately match regional differences, leading to regulation failures in high-demand areas. Regarding parking space reuse, in-depth analysis of idle periods after actual use misses high-value time slots available for reallocation, leaving some spaces vacant for extended periods and unable to respond to new requests. The matching and allocation process often aligns resources with requests on a one-to-one basis, ignoring the impact of time-of-day differences on matching effectiveness. This results in parking space allocation failures and invalid requests. Furthermore, a mechanism for calculating the supply-demand ratio has not been established, resulting in a failure to prioritize allocation targets in areas with limited parking resources, leading to scheduling imbalances and reduced utilization efficiency, further exacerbating parking conflicts and congestion in urban core areas. Summary of the Invention

[0005] The purpose of this invention is to solve the shortcomings of the existing technology and propose a smart city parking resource prediction and optimization decision-making platform based on big data.

[0006] To achieve the above objectives, the present invention adopts the following technical solutions: A big data-based smart city parking resource prediction and optimization decision-making platform includes: The parking spatiotemporal modeling module obtains the time and location data from the vehicle entry records in the target area, detects the spatial distance between adjacent vehicle parking points, combines the parking proximity determination threshold to determine the formation of parking segments, and generates parking segment distribution records; The demand potential identification module constructs a parking density sequence within the target area based on the parking section distribution records, determines whether the density exceeds a preset target area parking density classification standard, arranges parking section grade labels, and generates a parking demand grade list; The resource reuse prediction module calculates the difference between the maximum allowed stay time and the actual stay time of parking spaces in the parking section based on the parking demand level list, compares it with the parking space idle time threshold, selects eligible parking spaces and calculates the remaining time, establishes a correspondence between reusable parking spaces and requested time periods, and generates a reusable parking space prediction correspondence table; The parking space allocation decision module performs difference calculation based on the reusable time and vehicle request time in each docking relationship in the reusable parking space prediction correspondence table, determines whether it falls within the time tolerance range, screens candidate allocation items, sets the parking section push priority flag, and establishes a parking space allocation table to be pushed.

[0007] As a further solution of the present invention, the parking section distribution record includes vehicle stay duration characteristics, vehicle stop frequency characteristics, and the spatial coverage range of the parking section; the parking demand level list includes high-density parking section identification, medium-density parking section identification, low-density parking section identification, and management strategy labels corresponding to different density sections; the reusable parking space prediction correspondence table includes reusable parking space numbers, parking space idle time periods, corresponding vehicle request numbers, and parking space available time windows; the to-be-pushed parking space allocation table includes candidate allocation parking space numbers, corresponding vehicle request numbers, parking space request matching scores, and push priority levels.

[0008] As a further solution of the present invention, the parking spatiotemporal modeling module includes: The vehicle information extraction submodule extracts the time and location information of each vehicle based on the vehicle entry records in the target area, organizes the parking points of each vehicle in chronological order, calculates the parking positions between two adjacent vehicles, and generates a set of parking point pairs; The proximity distance judgment submodule calculates the Euclidean distance between any two sets of adjacent points in the parking point pair set based on the spatial coordinate data, compares the distance between each set of points with the minimum parking spacing standard, calculates the deviation between the actual spacing between adjacent parking spaces and the standard threshold, selects continuous parking point segments that meet the spatial proximity threshold condition, and generates a continuous parking point sequence; The segment attribute extraction submodule calculates the difference in start and end time of all vehicles in each segment based on the vehicle data corresponding to each point in the continuous parking point sequence, calculates the duration of the stop, and calculates the stop frequency by counting the number of vehicles in each segment. It then extracts the maximum and minimum boundaries based on the coordinate intervals of all parking points and converts them into spatial ranges to comprehensively generate parking segment distribution records.

[0009] As a further solution of the present invention, the demand potential identification module includes: The parking density construction submodule obtains the number of parking spots per unit area based on the spatial range data and stop frequency data in the parking segment distribution record. It then constructs a parking spot density sequence for each block based on the time window, combining the distribution of the parking spots in the regional grid and the corresponding time period, to obtain a parking density sequence value. The density level judgment submodule classifies the parking spaces according to the parking density sequence values ​​and the parking space density classification standard, based on the relationship between the density value and the standard interval position, calculates the density deviation measure of each parking section, and classifies the density level according to the interval, generating a parking density level label; The level list generation submodule sorts the parking density level labels in descending order, organizes the parking segment numbers and level information corresponding to each time window, merges the parking level data sequences within the area, and establishes a parking demand level list.

[0010] As a further solution of the present invention, the resource reuse prediction module includes: The dwell time difference calculation submodule collects the vehicle dwell time records for each parking space based on the parking section number in the parking demand level list, obtains the maximum allowed dwell time and the actual dwell time data, calculates the time difference between the two, and obtains the parking space dwell time difference data; The idle judgment and screening submodule obtains an acceptable waiting time threshold based on the parking space stay time difference data, makes a judgment based on the average idle time of vehicles leaving the corresponding parking section, screens parking spaces whose difference exceeds the idle time threshold, counts the remaining time by parking space number, calculates the estimated reuse remaining time of each qualified parking space, and merges and summarizes them by parking space to generate the reuse remaining time quantity; The time matching mapping submodule retrieves the vehicle request records in the concentrated distribution section based on the remaining reuse time, compares the remaining available time period of the parking space with the vehicle request time requirement, determines whether the usage interval overlap condition is met, performs time mapping between each parking space and the qualified request, and establishes a corresponding table for the prediction of reusable parking spaces.

[0011] As a further solution of the present invention, the parking space allocation decision module includes: The tolerance matching judgment submodule calculates the time difference based on the reusable duration and vehicle request duration data of each docking relationship in the reusable parking space prediction corresponding table, determines whether the difference is within the time tolerance range, screens the matching relationships between the qualified parking spaces and requests, and obtains a set of candidate allocation relationships; The candidate ratio calculation submodule classifies and counts the candidate allocation relationship set by parking section. In each parking section, the number of candidate allocation spaces and the corresponding number of vehicle requests are extracted. Based on the basic allocation standard set for parking supply adequacy in municipal planning, the ratio of the two is compared to see if it exceeds the matching benchmark threshold. Supply-demand ratio data for each parking section is then obtained to generate an allocation supply-demand ratio sequence. The priority push marking submodule selects parking sections with supply-demand ratio values ​​higher than the set benchmark ratio according to the allocation supply-demand ratio sequence, sets the push priority flag, organizes the pushable objects and their allocation numbers in combination with the corresponding candidate allocation relationship information, and establishes a parking space allocation table to be pushed.

[0012] As a further embodiment of the present invention, the system further comprises: The trend feedback optimization module extracts the response rate, usage completion rate, and response delay time of each allocated parking space based on the parking space allocation table to be pushed, determines the adaptability of each parking section allocation mode, and combines the multi-section solution structure based on the parking density trend in the subsequent time period to generate a parking resource optimization management plan; The parking resource optimization management plan includes statistical records of parking space allocation response rates in each section, parking space utilization completion rate analysis results, allocation response delay evaluation results, and multi-section allocation combination structure.

[0013] As a further solution of the present invention, the trend feedback optimization module includes: The response data extraction submodule extracts the response rate data, usage completion rate data, and response delay time data of each parking space based on the parking space allocation table to be pushed, obtains parking space response records, parking payment confirmation information, and user operation timestamps, performs data cross-validation, and integrates them into a time series format to obtain a deployment response indicator sequence; The deployment adaptation determination submodule normalizes the response rate, usage completion rate, and response delay time of each parking section based on the deployment response indicator sequence, calculates the difference between the response rate and the response time limit standard, calculates the average matching degree according to the indicator dimension, determines the adaptability of the deployment mode of each section, and obtains the deployment adaptation analysis results; The optimization combination generation submodule divides the parking sections into groups according to the allocation fitness analysis results and the parking density trend information in the subsequent time period, sets the combination structure matching, constructs the parking resource allocation logic relationship, and establishes the parking resource optimization management plan.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, by extracting the spatiotemporal behavior of vehicles and determining continuous parking spots, high-frequency parking sections are divided to improve the recognition accuracy of micro-parking distribution; a parking density sequence is constructed and density level division is introduced to achieve differentiated characterization and dynamic classification of regional parking needs; by combining the comparison of the maximum allowed stay time with the actual occupancy time, reusable time period parking space resources are identified, and the reuse potential of non-occupied parking spaces is explored; parking spaces and requests are calibrated and matched based on the duration difference and time tolerance to enhance scheduling accuracy and response timeliness; a supply-demand ratio push priority strategy is introduced to tilt resources towards high-demand areas, improve the target adaptability of the allocation strategy, and overall enhance the precision and timeliness of urban parking resource allocation. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a platform flow chart of the present invention; Figure 2 This is a flow chart of the parking spatiotemporal modeling module of the present invention; Figure 3 This is a flow chart of the demand potential identification module of the present invention; Figure 4 This is a flow chart of the resource reuse prediction module of the present invention; Figure 5 This is a flow chart of the parking space allocation decision module of the present invention; Figure 6 This is the flow chart of the trend feedback optimization module of the present invention. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0017] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0018] See also Figure 1 , a smart city parking resource prediction and optimization decision-making platform based on big data includes: The parking spatiotemporal modeling module obtains the time and location data from vehicle entry records in the target area, sequentially detects the spatial distances between adjacent parking spots, and determines whether several consecutive groups of spatial distances are all below the parking proximity determination threshold (based on the minimum parking spacing standard (usually 5-8 meters) specified in the "Urban Road Parking Space Setup Specifications" to determine whether adjacent parking spots constitute a continuous parking section). The module then counts the parking spots that meet the conditions and forms parking sections. It then extracts the vehicle dwell time, dwell frequency, and spatial extent of each parking section to generate a parking section distribution record. The demand potential identification module constructs a parking density sequence within the target area based on the spatial extent and dwell frequency of parking segment distribution records. It then determines whether the density exceeds the preset target area parking density classification standard (based on the "Urban Static Traffic Assessment Guidelines" issued by the traffic management department, which classifies parking spaces into three density levels: high, medium, and low, based on the number of parking spots per unit area, with each level corresponding to a different management strategy). The module then categorizes and labels parking spaces by density, arranges the parking segment classification labels within each time window, and generates a parking demand level list. The resource reuse prediction module collects vehicle dwell time records for parking spaces within a parking section based on a parking demand level list, calculates the difference between the maximum permitted dwell time and the actual dwell time, and determines whether the difference exceeds the parking space idle time threshold (using the acceptable waiting time parameter from traffic engineering (usually set at 15 minutes) and the average length of time a parking space remains idle after a vehicle leaves the parking lot, obtained through historical data statistics). The module then selects eligible parking spaces and calculates the remaining time. Based on the centralized distribution and the usage time requirements in vehicle requests, it establishes a time period correspondence between reusable parking spaces and requests, generating a reusable parking space prediction table. The parking space allocation decision module calculates the difference between the reusable duration and the vehicle request duration of each docking relationship in the reusable parking space prediction table, and determines whether it falls within the time tolerance range (calculated based on GPS positioning accuracy and road network travel time according to the ISO15392 intelligent transportation system standard (usually set to ±5 minutes)). It selects the relationships that meet the requirements and marks them as candidate allocation items. It calculates the ratio of the number of candidate allocation spaces to the number of requests in each parking section. For parking sections with a ratio greater than the supply-demand matching benchmark ratio (using the parking supply adequacy ratio indicator in municipal planning, with a parking space / request ratio of 1:1.2 as the basic allocation standard), a push priority flag is set, and a parking space allocation table to be pushed is established. Based on the table of parking space allocations to be pushed, the trend feedback optimization module extracts the response rate, usage completion rate (the percentage of allocated parking orders that are actually completed, obtained through cross-validation between the electronic payment system and geomagnetic sensor data) and response delay time (the time interval from receiving the allocation request to user confirmation of the operation, which should not exceed 60 seconds in accordance with the "Smart City Information Service Response Standard") of each allocated parking space. It determines the adaptability of the allocation mode of each parking section, combines the parking density trend in subsequent time periods, and generates a multi-segment solution structure to optimize parking resource management.

[0019] The parking section distribution records include the characteristics of vehicle stay time, vehicle stay frequency, and the spatial coverage of the parking section. The parking demand level list includes high-density parking section identification, medium-density parking section identification, low-density parking section identification, and management strategy labels corresponding to different density sections. The reusable parking space prediction correspondence table includes the reusable parking space number, parking space idle time period, corresponding vehicle request number, and parking space available time window. The parking space allocation table to be pushed includes the candidate allocation parking space number, corresponding vehicle request number, parking space request matching score, and push priority level. The parking resource optimization management plan includes statistical records of parking space allocation response rates in each section, parking space utilization completion rate analysis results, allocation response delay evaluation results, and multi-section allocation combination structure.

[0020] See also Figure 2 , the parking spatiotemporal modeling module includes: The vehicle information extraction submodule extracts the time and location information of each vehicle based on the vehicle entry records in the target area, organizes the parking points of each vehicle in chronological order, calculates the parking positions between two adjacent vehicles, and generates a set of parking point pairs; Based on the vehicle entry records in the target area, the data format of the entry records is first standardized, and the record time field and geographic coordinate field are extracted. By constructing a data structure, the entry time and parking location of each vehicle are unified into timestamp and latitude and longitude values. The structure is then stored as a vehicle information array. All vehicle information is further sorted in ascending order according to the timestamp field to ensure that subsequent processing can be performed based on time sequence. Next, consecutive vehicle records are selected and combined into a point pair set. Each pair of points contains four basic fields: start time, end time, start location, and end location. For example, vehicle A enters at location (116.397, 39.908) at 08:00 on April 10, 2024, and vehicle B enters at (116.399, 39.9085) at 08:03. After time sorting, the two form a point pair, where the first item is vehicle A and the second item is vehicle B. Their spatial locations are extracted as a set of parking point coordinates, and their respective entry times are recorded for subsequent frequency and duration statistics. Each point pair is input as a record in the list. Table 1 Initial data table of parking points As shown in Table 1, by combining time sorting and spatial point pairs, we can construct point pair data with time and space dimensions, providing the necessary basis for subsequent spatial distance judgment and segment generation, and ultimately obtaining a set of parking point pairs.

[0021] The proximity distance judgment submodule calculates the Euclidean distance between any two sets of adjacent points in the parking point pair set based on their spatial coordinate data, and compares them with the minimum parking spacing standard using the formula: ; Calculate the deviation between the actual spacing between adjacent berths and the standard threshold , filter the continuous parking point segments that meet the spatial proximity threshold conditions and generate a continuous parking point sequence, where, Indicates the The horizontal coordinate of the parking point, Indicates the The vertical coordinate of the parking point, Indicates the The horizontal coordinate of the parking point, Indicates the The vertical coordinate of the parking point, represents the total number of adjacent point pairs, It represents the parking proximity determination threshold, which is derived from the minimum parking spacing standard in the "Urban Road Parking Space Setting Specifications". Indicates the The parking frequency of the vehicle corresponding to each parking point, Indicates the The parking time of the vehicle corresponding to each parking point, Indicates the The number of vehicles that have been determined to be adjacent in a continuous point segment; Based on each set of adjacent coordinate data in the parking point pair set, the longitude and latitude differences between the starting and ending positions in each point pair are first calculated, and the spatial distance is calculated using the Euclidean distance calculation method. The formula is as follows: ; in, For the The starting point coordinates of the point pair, For the The starting point coordinates of the point pairs are in meters. Then all distance values ​​are compared with the parking proximity determination threshold. Compare, if continuous The distance values ​​satisfy , then the section is determined to be a continuous parking section. In the judgment example, the parking proximity determination threshold is set to 7 meters, which is derived from the "Urban Road Parking Space Setting Specifications" which stipulates that the minimum parking distance is 5 to 8 meters, and the median value is used as the calculation standard.

[0022] Combined with actual data, if the spatial distances of the three sets of point pairs are: Group 1: ; Group 2: ; Group 3: ; Since the above distances all exceed the 7-meter threshold, the current point pair group does not meet the judgment conditions and does not constitute a continuous segment; If the sample data is adjusted to adjacent point distances of 4.2 meters, 5.1 meters, and 4.8 meters, then since all three are less than 7 meters, the number of consecutive point segments is Meet the judgment conditions; introduce frequency and stay time for composite judgment, combined with the frequency of the vehicle at each point in the section , length of stay Minutes, and the number of vehicles in the current segment , enter the following formula to calculate the composite index: ; Substituting the example values ​​into the calculation results is: Item 1: ; Item 2: ; Item 3: ; but: ; In order to distinguish whether the behavior patterns between vehicles are consistent, the behavior difference index is used The value is used as a standard, and the fluctuation of the frequency and length of stay of vehicles in the same parking section due to factors such as morning and evening peaks, weekdays / holidays, etc. is referred to. The threshold value 10 is set as the upper limit of "strong behavioral consistency", representing a combination of vehicles with a frequency difference of no more than 1 time and a stay time difference of no more than 3 minutes. Less than 10; the threshold of 20 is based on the combination of frequency exceeding 2 times and the difference in stay time exceeding 6 minutes, indicating that the behavior pattern is inconsistent and the deviation is large, so it corresponds to The results will be eliminated. This value range is set based on the differences in the behavior amplitude of vehicle numerical samples in each group, and is slightly adjusted with the changes in the density of regional road sections. It is an adjustable parameter.

[0023] like : It is believed that spatial continuity and behavioral consistency are significant, constituting an effective continuous segment; like : Partial matching segment, which needs to be judged in combination with the context; like : It is considered that the behavior consistency is not met and it is determined to be an invalid continuous segment.

[0024] The current calculated value is 21.13, which exceeds the upper limit of the judgment range. Therefore, although the spatial distance of this point segment meets the requirements, the vehicle's stop frequency and time vary greatly, and it does not constitute a continuous parking point segment. It will not be included in the processing range of the subsequent "Segment Attribute Extraction" step.

[0025] The operational logic of this formula mainly revolves around the difference between spatial distance and behavioral characteristics. First, the latitude and longitude differences of each set of adjacent parking points are squared and added together to obtain the squared spatial Euclidean distance between the point pairs. The squared value is then divided by the parking proximity determination threshold. Standardization is achieved to reflect the relative degree of the current point-to-space distance within the preset distance range; then, the behavioral parameter of each parking point - the frequency of stay and length of stay The sum of the squares is used to represent the intensity of the vehicle's behavior at that point, and the value is divided by the number of vehicles. Sparsity adjustment is introduced, and then the square root of the overall behavior intensity value is taken to compress its magnitude and smooth the scale of behavioral differences; finally, the spatial term is subtracted from the behavioral term, and the absolute value is taken to ensure that the direction of the difference does not affect the measurement. Then all point pair terms are summed to summarize the space-behavior matching deviation of the entire segment. Therefore, the addition in the operation reflects the aggregation of indicators within the spatial or behavioral dimension, the division and square root realize normalization and regulation amplitude, the absolute value ensures that the difference is positive and comparable, and the summation operation reflects the overall consistency trend between multiple point pairs.

[0026] The segment attribute extraction submodule uses the vehicle data corresponding to each point in the continuous parking point sequence to count the start and end time differences of all vehicles in each segment, calculate the dwell time, and count the number of vehicles in each segment to obtain the dwell frequency. Based on the coordinate intervals of all parking points, the maximum and minimum boundaries are extracted and converted into spatial ranges to comprehensively generate parking segment distribution records. Based on the vehicle number, entry time and spatial coordinate data of all parking points in the continuous parking point sequence, we first extract the entry time and exit time of each vehicle in the current segment, and calculate the time difference to obtain the single vehicle's stay time. Assuming that vehicle X stays in the segment from 08:10 to 08:24, the stay time is 14 minutes. Then, we calculate the average stay time of all vehicles in the current segment. If there are 3 vehicles with stay times of 12 minutes, 14 minutes and 10 minutes respectively, the average within the segment is 12 minutes, which is used as the stay time indicator within the segment. Then, we calculate the number of vehicles appearing in the current segment, which is the stay frequency, which is 3 times here. The spatial range is constructed by the minimum and maximum longitude and latitude of the segment to form a longitude and latitude rectangle. Assuming the minimum point is (116.3970, 39.9080) and the maximum point is (116.4030, 39.9101), the spatial range is approximately: longitude span , latitude span , which is converted into actual meter units, is approximately 660 meters × 233 meters; finally, the dwell time, dwell frequency, and spatial coverage values ​​of all vehicles in the current segment are integrated and used as the characteristic representation items of the segment to generate the parking segment distribution record.

[0027] See also Figure 3 , the demand potential identification module includes: The parking density construction submodule uses the spatial range data and stop frequency data in the parking segment distribution record to obtain the number of parking spots per unit area. Combining the distribution of parking spots in the regional grid and the corresponding time period, it constructs a parking spot density sequence for each block according to the time window, and obtains the parking density sequence value. Based on the spatial range and stop frequency in the parking segment distribution records, the density construction process of the parking area first needs to extract the parking segment boundary coordinates in each time window and calculate the corresponding area value based on the boundary. For example, the polygonal area formed by the parking segment boundary of a certain area is 480 square meters. At the same time, the stop frequency data of all vehicles in the area within the time window are counted. If the area contains 12 parking spots, the frequency corresponding to each spot is times, the average frequency is 2.5 times. Next, the parking density per unit area is calculated by calculating the ratio of the number of vehicle parking points to the area. That is, the parking density value is points / square meter, and then perform the above operation on the parking areas within all time windows to construct a parking density sequence for different areas within the corresponding time period. To achieve repeatability, the following table lists the calculation results of the number and area of ​​parking spots in the example area under each time period, as shown in Table 2.

[0028] Table 2 Parking area density calculation table As shown in Table 2, under different time windows, the hourly parking density sequence values ​​are calculated by the ratio of the area to the number of parking spots. For example, the parking density during the 09:00-10:00 period is points / square meter, which is consistent with the 08:00-09:00 period, while 10:00-11:00 is Points / square meter, the parking density sequence value will be used as the input basis for subsequent density level determination to obtain the parking density sequence value.

[0029] The density level judgment submodule classifies parking spaces according to the parking density sequence value and the parking space density classification standard, and the relationship between the density value and the standard interval position, using the formula: ; Calculate the density deviation metric for each parking segment , and classify the density level according to the interval, and generate parking density level labels, where Indicates the The parking density value of each parking block, Indicates the standard boundary value of the parking section corresponding to the density level. 、 Respectively represent the area scale coefficients of the parking section in the horizontal and vertical directions, Indicates the frequency of vehicle stops in the parking section. Indicates the total number of parking blocks involved in the calculation; The parking area density level is determined based on the parking density sequence values ​​and the parking density grading standards in the static traffic assessment guide. First, the deviation between the density value of each area in the parking density sequence and the set grading standard is calculated. The grading standard is set as a high density range greater than 0.03 points / square meter, a medium density range of 0.015–0.03 points / square meter, and a low density range of less than 0.015 points / square meter. The parking density values ​​of three example areas are selected as 0.025, 0.020, and 0.012 points / square meter, respectively, and are substituted into the following formula for determination: ; Setting parameters: ; ; ; ; ; Substituting the above values ​​into the calculation: ; The results show that the parking density deviation measure within the selected time window is 0.0419, which falls within the medium density range. The system generates a parking density grade label based on this. The beneficial formula is that it calculates the average value by combining the normalized deviation between the parking density value and the standard boundary value with the product of the parking frequency, so that density judgment is no longer limited to the judgment of absolute density, but introduces frequency weighting to achieve multi-factor graded evaluation.

[0030] The operational logic of this formula is to comprehensively evaluate the degree of deviation between the density value of each parking block and its density grade standard, and introduce spatial scale and parking frequency as adjustment factors to enhance the representativeness of the evaluation results. Indicates the difference between the actual parking density value of each block and its standard level boundary value, which is used to characterize the degree of deviation. Characterize the spatial scale influencing factors of the block. By taking the square root of the sum of the horizontal and vertical scales plus one, a unified measurement of parking density at different regional scales is achieved to prevent the regional size from causing an uneven impact on the deviation value. The product term is added with This indicates that the frequency of vehicle stops within a block is used as an evaluation weight. A higher frequency indicates a more active block in traffic use, and deviations should be given greater attention. Therefore, the three factors are combined and averaged to form a normalized deviation index across multiple blocks. Absolute value calculations prevent positive and negative offsets and preserve the absolute intensity of deviation. This logic, through the structure of "difference × weight / scale adjustment," reflects the multi-factor approach to calculating deviation levels in the context of spatial scale inconsistency.

[0031] The level list generation submodule sorts parking density level labels from high to low, organizes parking segment numbers and level information corresponding to each time window, merges parking level data sequences within the area, and creates a parking demand level list; After obtaining the parking density level labels, the system sorts different time windows and spatial blocks from large to small according to the density deviation, organizes the numbers corresponding to each parking section under each time window and its density level, and then classifies and labels them. For example, number A01 corresponds to the medium-density area, number A02 corresponds to the low-density area, and number A03 corresponds to the medium-density area. The level list of parking areas within the time period is completed, and the correspondence between each set of time windows and parking numbers is obtained. For example, in the time period of 08:00-09:00, the density level of the A01 parking section is medium-density, A02 is low-density, and A03 is medium-density. Records are sorted and lists are generated accordingly, and finally a parking demand level list is established.

[0032] See also Figure 4 ,The resource reuse prediction module includes: The dwell time difference calculation submodule collects the vehicle dwell time records for each parking space based on the parking section number in the parking demand level list, obtains the maximum allowed dwell time and the actual dwell time data, calculates the time difference between the two, and obtains the parking space dwell time difference data; Based on the parking segment numbers listed in the parking demand level list, vehicle dwell time records for all parking spaces in the area are collected from the database. The collected fields include each parking space number and vehicle entry and exit times. This is used to calculate the actual dwell time for each parking space. Simultaneously, the maximum dwell time threshold for parking spaces set in the management system is retrieved. This threshold is set based on regional characteristics, for example, 120 minutes in residential areas and 60 minutes in commercial areas. A matching operation is performed based on parking distribution characteristics. For example, parking space P001 has a maximum dwell time of 120 minutes and an actual dwell time of 95 minutes. By subtracting the two, the difference is 25 minutes. Similarly, the difference for parking space P002 is −15 minutes, indicating an overstay. This information will not be included in the subsequent reuse period calculation range. Parking space numbers with valid differences are compiled into a dataset, which serves as the data basis for subsequent parking space reuse analysis and generates parking space dwell time difference values.

[0033] Table 3 Reusable parking space related parameters

[0034] As shown in Table 3, the core parameters involved in the subsequent calculations are listed, including the maximum allowed stay time, actual stay time, request frequency, and spatial location parameters.

[0035] The idle judgment and screening submodule obtains the acceptable waiting time threshold based on the parking space stay time difference data, and makes a judgment based on the average idle time of vehicles leaving the parking section. It screens out parking spaces with a difference exceeding the idle time threshold, and calculates the remaining time by parking space number using the formula: ; Calculate the estimated remaining reuse time for each eligible parking space , and merge and summarize by parking space to generate the remaining reuse time, where Indicates the The maximum allowed stay time for each parking space, Indicates the actual stay time, Indicates the demand level value of the sub-block where the parking space is located. Indicates the number of vehicle requests for parking spaces within the time window. Indicates the average time difference between the parking space and the nearest request point. Indicates the length of the parking space idle period, Indicates the number of parking spaces that meet the requirements in the current section; Based on the parking space stay time difference obtained above, the benchmark value for judging whether the parking space is idle is set to 15 minutes. That is, when the difference is greater than this value, the parking space is judged to have a reusable period. Subsequently, the number of vehicle request records in the area where the corresponding parking space is located, the time difference of the most recent request point, the length of the idle time, and the demand level value of the corresponding sub-block are retrieved, and each type of data is read and collected item by item. The data items required for the calculation formula are sorted according to the parking space number and substituted into the formula for calculation.

[0036] Taking parking space P001 as an example, the maximum allowed stay time is , actual stay time , requirement level value , number of requests , request point time difference , length of idle period , put it into the formula and we get: ; This value indicates that the remaining reuse time of the parking space numbered P001 is 33.921 minutes. Then all eligible parking spaces are calculated and summarized one by one in the same way to obtain the remaining reuse time in the current section.

[0037] The calculation logic of this formula is to comprehensively evaluate the reusable time of parking spaces. The combination relationship between the parameters reflects the multi-dimensional quantification of time difference, demand level, usage frequency and spatial association. The first item Used to measure the match between the available time of a parking space and the demand level of the area where it is located. The absolute value ensures that the time difference is a positive number and is multiplied by the demand level value. In areas with higher demand levels, idle time is more important, and the Indicates that the request frequency is used as a suppression term. When a parking space has a large number of requests, its reuse feasibility is limited. Therefore, a nonlinear reduction is achieved through the square root function. The second term It reflects the spatial and temporal proximity between the parking space and the request. The numerator is the time difference from the nearest request point, and the denominator is the length of the idle period plus one, which is used to avoid the denominator being zero and at the same time weakening the length of the idle interval. The entire formula combines the two terms into one through addition, taking into account both local idle capacity and its possibility of serving surrounding requests, and constructing a reuse estimation structure that integrates time redundancy, spatial efficiency, and demand urgency. The time matching mapping submodule retrieves the vehicle request records in the concentrated distribution section based on the remaining reuse time, compares the remaining available time period of the parking space with the vehicle request time requirement, determines whether the use interval overlap condition is met, performs time mapping between each parking space and the qualified request, and establishes a corresponding table for the prediction of reusable parking spaces; According to the obtained data on the remaining reuse time, the vehicle request records in the concentrated distribution area are retrieved, the request number and the corresponding time period information are obtained, and the remaining available time period of each parking space is compared with the required time period of the vehicle request to determine whether there is an overlap between the time periods. If there is an overlap, it is considered that the current request can match the parking space. For example, if the available time period of the P001 parking space is from 15:00 to 15:25 on April 23, 2025, and the time period of a certain request is from 15:10 to 15:20, then the two intervals meet the coverage condition and a mapping relationship can be established. Finally, a number mapping is performed on all successfully matched parking spaces and request numbers to establish a corresponding table for the prediction of reusable parking spaces.

[0038] See also Figure 5 ,The parking space allocation decision module includes: The tolerance matching judgment submodule calculates the time difference based on the reusable duration and vehicle request duration data of each docking relationship in the reusable parking space prediction corresponding table, determines whether the difference is within the time tolerance range, screens the matching relationships between the qualified parking spaces and requests, and obtains the candidate allocation relationship set; Obtain the correspondence between parking spaces and vehicle requests in the reusable parking space prediction table. First, extract the reusable parking space number, the remaining time for reusability, and the corresponding vehicle request number and request time length recorded in each data item. Unify the above time data numerically and convert them into integers by minute. Then calculate the time difference between the two. For example, the remaining time for reusability in the parking space record numbered A01 is 55 minutes, and the corresponding request time is 50 minutes, so the difference is 5 minutes. According to the time tolerance definition in the ISO15392 intelligent transportation system standard, the standard recommends using ± 5 minutes is the acceptable error limit under the influence of spatiotemporal error, and the difference is within the tolerance range. Therefore, parking space A01 is marked as a matching relationship. For example, parking space A03 has a remaining reuse time of 35 minutes and a vehicle request of 40 minutes, with a difference of -5 minutes. Although the difference is 5 minutes, the direction is insufficient. Therefore, it is necessary to determine whether this type of reverse error is allowed based on system requirements. If only non-negative differences are allowed in the configuration file, the relationship will not be selected into the matching set. Then, all corresponding relationships are screened according to this rule, and the matching items that meet the conditions are sorted into a candidate allocation relationship set, as shown in Table 4.

[0039] Table 4 Reusable parking space matching relationship table

[0040] As shown in Table 4, the parking spaces in the parking sections numbered A01 and A02 both meet the matching condition between the reusable duration and the request time, constituting a candidate allocation relationship set.

[0041] The candidate ratio calculation submodule classifies and counts the candidate allocation relationship set by parking section. It extracts the number of candidate allocation spaces and the corresponding number of vehicle requests in each parking section. Based on the basic allocation standard set for parking supply adequacy in municipal planning, it compares the ratio of the two to see if it exceeds the matching benchmark threshold. It then obtains the supply-demand ratio data for each parking section and generates an allocation supply-demand ratio sequence. According to the records in the candidate allocation relationship set, they are grouped and counted by parking segment number. The total number of candidate parking spaces and the total number of corresponding requests are recorded in each group, and the ratio is calculated. For example, in Table 1, there are 36 candidate parking spaces in segment A01 and 30 requests, with a supply-demand ratio of 36:30, or 1.2. Similarly, the ratio of segment A02 is 28:25, or 1.12. Although segment A03 has records of parking spaces and requests, it is not included in the candidate set because it does not meet the tolerance conditions. Therefore, it is not included in this step of statistics, and the benchmark matching ratio is then set. According to the municipal parking facility planning guidelines, the value is recommended to be set to 1:1.2, which means that the parking space should have at least 20% redundancy for the request before it can be pushed. Therefore, the calculated ratio of each section is compared with 1.2. When the section ratio is greater than or equal to 1.2, it is determined that its parking space resources have the basic conditions for pushing. Otherwise, it is determined that the supply and demand ratio does not meet the standard requirements. For example, the A01 section ratio is 1.2, which meets the conditions, and the A02 ratio is 1.12, which does not meet the conditions. Finally, all calculation results are summarized to form a supply and demand ratio sequence.

[0042] The priority push marking submodule selects parking sections with supply-demand ratios higher than the set benchmark ratio based on the allocation supply-demand ratio sequence, sets push priority flags, organizes pushable objects and their allocation numbers based on the corresponding candidate allocation relationship information, and establishes a parking space allocation table to be pushed; According to the ratio judgment result of each parking section in the allocation supply and demand ratio sequence, the sections with ratios greater than or equal to the benchmark value are screened out, and a push priority flag is set for them. In actual operation, the priority identification code can be set according to the section number. For example, the A01 section number that meets the conditions is set to "push priority = 1", and the section number that does not meet the conditions is set to "push priority = 0". Subsequently, the parking space number and matching request information of the corresponding section are called from the candidate allocation relationship set to form the basic record structure of the push plan table. Each row contains the parking space number, the corresponding request number, the time period to be allocated and the priority mark, thereby constructing the parking space allocation table to be pushed, which serves as the input content for the subsequent system allocation plan generation.

[0043] See also Figure 6 , the trend feedback optimization module includes: The response data extraction submodule extracts the response rate data, usage completion rate data, and response delay time data for each parking space based on the parking space allocation table to be pushed. It obtains parking space response records, parking payment confirmation information, and user operation timestamps, performs data cross-validation, and integrates them into a time series format to obtain the allocation response indicator sequence. Based on the parking space allocation table to be pushed, indicators such as the response rate, usage completion rate, and response delay time of each parking space in actual application are obtained. First, the task execution record of each assigned parking space is retrieved, the response timestamp and confirmation timestamp are extracted, and their time interval is calculated. For example, the task allocation time of a parking space is 10:00:00, and the user clicks to confirm at 10:00:38, then the response delay time of the parking space is 38 seconds. Then, the total number of response records of each parking space and the number of responded tasks are summarized, and the formula response rate = number of responded tasks ÷ total number of tasks × 100% is used for calculation. For example, if parking space P101 is assigned 12 tasks on the same day, and 11 of them are responded, the response rate is 11 ÷ 12 × 100% = 91.67%. Secondly, the completion records are cross-extracted from the geomagnetic sensor data and the electronic payment system, and the number of tasks in which payment is completed and actual parking behavior occurs is counted. Combined with the total number of responses, the response rate is calculated. Completion rate = actual number of completions ÷ number of responses × 100%. If P101 has 11 responses, 8 of which were followed by confirmation and payment, the completion rate is 8 ÷ 11 × 100% = 72.73%. The collected data results are shown in Table 1. The response delay time is determined to determine whether it meets the 60-second threshold set by the "Smart City Information Service Response Standard". Records with delays greater than 60 seconds are classified as "unqualified". For example, the response delay of parking space P202 is 65 seconds, which is considered to exceed the standard. At the same time, according to the data in Table 1, it can be concluded that parking spaces with a response rate above 85% are classified as high-response segments, while those below 80% are classified as low-response segments. The completion rate is divided into adaptation levels with 75% as the cutoff point. Finally, the above three indicators are combined with the parking segment to which each parking space belongs for classification and integration to form a segment-level allocation response indicator sequence, which serves as the data basis for subsequent optimization analysis.

[0044] Table 5 Parking space response data table

[0045] As shown in Table 5, the response delay time of parking space P202 has exceeded the standard upper limit of 60 seconds, while the response and completion rates of P101 and P303 can be classified into different response levels to further participate in the deployment adaptability judgment.

[0046] The deployment adaptation determination submodule normalizes the response rate, usage completion rate, and response delay time of each parking section based on the deployment response indicator sequence, calculates the difference between the response rate and the response time limit standard, calculates the average matching degree according to the indicator dimension, determines the adaptability of the deployment mode of each section, and obtains the deployment adaptation analysis results; Based on the deployment response indicator sequence, the response rate, completion rate and response delay time of each parking section are first normalized respectively. The normalization adopts the minimum-maximum normalization method. For example, for the response rate normalization, the maximum value is set to 100% and the minimum value is 70%. The normalized response rate of parking space P101 is (92.5-70) ÷ (100-70) = 0.75. Similarly, the completion rate is set to 90% and the minimum value is 60%. The completion rate of P101 is 78.4%, and the normalized value is (78.4-60) ÷ (90-60) = 0.61. The response delay is based on the maximum 60 seconds as the standard line, and the normalized value is calculated in reverse, that is, the normalized value = 1-actual delay ÷ 60. The response delay of P101 is 42 seconds, and the normalized value is 1-42÷60=0.30. The three normalized values ​​are weighted averaged according to the weights of response rate 0.4, completion rate 0.4, and response delay 0.2, and the deployment fitness value is 0.75×0.4+0.61×0.4+0.30×0.2=0.63. It is further classified according to the fitness range, with 0.8 and above as the high adaptation zone, 0.6 to 0.8 as the medium adaptation zone, and below 0.6 as the low adaptation zone. If the fitness values ​​of most parking spaces in a section fall above 0.8, the entire section can be marked as a high adaptation section. In actual operation, all parking spaces need to be traversed and classified and aggregated by section to finally generate the deployment fitness analysis results for combination strategy analysis.

[0047] The optimization combination generation submodule divides parking sections into groups based on the results of the allocation fitness analysis and the parking density trend information in the subsequent time period, sets the combination structure matching, constructs the parking resource allocation logic relationship, and establishes a parking resource optimization management plan; According to the results of the deployment fitness analysis, combined with the parking density trend data in the subsequent time period, the high-density section number and its time period are obtained. The density is obtained based on the historical vehicle entry frequency statistics. The combination strategy structure is set. Strategy one is "high-density section priority deployment", strategy two is "high fitness + rising trend", and strategy three is "full area joint coverage". A combination score is set for each parking section. The scoring model is determined by the deployment fitness value and the trend increase value. For example, if the deployment fitness of a section is 0.78 and the trend increase is +20, the combination score is set to 0.78+0.02×trend increase=0.78+0.40=1.18. The combination threshold is set to 1.1. The satisfied segments are marked as first-level combination segments and identified as combinable units in the structure mapping matrix. The parking space allocation rule mapping format is generated through the combination marking structure. For example, Z01-T2-C1 indicates that the Z01 segment belongs to the first-level allocation combination in the T2 time window and is associated with the C1 structure strategy. Finally, all combination attribute segments are sorted according to the time window, and a mapping table and allocation command structure are established to form a complete parking resource optimization management plan.

[0048] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A big data-based smart city parking resource prediction and optimization decision-making platform, characterized by: The platform includes: The parking spatiotemporal modeling module obtains the time and location data from the vehicle entry records in the target area, detects the spatial distance between adjacent vehicle parking points, combines the parking proximity determination threshold to determine the formation of parking segments, and generates parking segment distribution records; The demand potential identification module constructs a parking density sequence within the target area based on the parking section distribution records, determines whether the density exceeds a preset target area parking density classification standard, arranges parking section grade labels, and generates a parking demand grade list; The resource reuse prediction module calculates the difference between the maximum allowed stay time and the actual stay time of parking spaces in the parking section based on the parking demand level list, compares it with the parking space idle time threshold, selects eligible parking spaces and calculates the remaining time, establishes a correspondence between reusable parking spaces and requested time periods, and generates a reusable parking space prediction correspondence table; The parking space allocation decision module performs difference calculation based on the reusable time and vehicle request time in each docking relationship in the reusable parking space prediction correspondence table, determines whether it falls within the time tolerance range, screens candidate allocation items, sets the parking section push priority flag, and establishes a parking space allocation table to be pushed.

2. The big data-based smart city parking resource prediction and optimization decision-making platform according to claim 1 is characterized in that: The parking section distribution record includes vehicle stay duration characteristics, vehicle stay frequency characteristics, and the spatial coverage of the parking section. The parking demand level list includes high-density parking section identification, medium-density parking section identification, low-density parking section identification, and management strategy labels corresponding to different density sections. The reusable parking space prediction correspondence table includes reusable parking space numbers, parking space idle time periods, corresponding vehicle request numbers, and parking space available time windows. The to-be-pushed parking space allocation table includes candidate allocation parking space numbers, corresponding vehicle request numbers, parking space request matching scores, and push priority levels.

3. The big data-based smart city parking resource prediction and optimization decision-making platform according to claim 1 is characterized in that: The parking spatiotemporal modeling module includes: The vehicle information extraction submodule extracts the time and location information of each vehicle based on the vehicle entry records in the target area, organizes the parking points of each vehicle in chronological order, calculates the parking positions between two adjacent vehicles, and generates a set of parking point pairs; The proximity distance judgment submodule calculates the Euclidean distance between any two sets of adjacent points in the parking point pair set based on the spatial coordinate data, compares the distance between each set of points with the minimum parking spacing standard, calculates the deviation between the actual spacing between adjacent parking spaces and the standard threshold, selects continuous parking point segments that meet the spatial proximity threshold condition, and generates a continuous parking point sequence; The segment attribute extraction submodule calculates the difference in start and end time of all vehicles in each segment based on the vehicle data corresponding to each point in the continuous parking point sequence, calculates the duration of the stop, and calculates the stop frequency by counting the number of vehicles in each segment. It then extracts the maximum and minimum boundaries based on the coordinate intervals of all parking points and converts them into spatial ranges to comprehensively generate parking segment distribution records.

4. The big data-based smart city parking resource prediction and optimization decision-making platform according to claim 1 is characterized in that: The demand potential identification module includes: The parking density construction submodule obtains the number of parking spots per unit area based on the spatial range data and stop frequency data in the parking segment distribution record. It then constructs a parking spot density sequence for each block based on the time window, combining the distribution of the parking spots in the regional grid and the corresponding time period, to obtain a parking density sequence value. The density level judgment submodule classifies the parking spaces according to the parking density sequence values ​​and the parking space density classification standard, based on the relationship between the density value and the standard interval position, calculates the density deviation measure of each parking section, and classifies the density level according to the interval, generating a parking density level label; The level list generation submodule sorts the parking density level labels in descending order, organizes the parking segment numbers and level information corresponding to each time window, merges the parking level data sequences within the area, and establishes a parking demand level list.

5. The big data-based smart city parking resource prediction and optimization decision-making platform according to claim 1 is characterized in that: The resource reuse prediction module includes: The dwell time difference calculation submodule collects the vehicle dwell time records for each parking space based on the parking section number in the parking demand level list, obtains the maximum allowed dwell time and the actual dwell time data, calculates the time difference between the two, and obtains the parking space dwell time difference data; The idle judgment and screening submodule obtains an acceptable waiting time threshold based on the parking space stay time difference data, makes a judgment based on the average idle time of vehicles leaving the corresponding parking section, screens parking spaces whose difference exceeds the idle time threshold, counts the remaining time by parking space number, calculates the estimated reuse remaining time of each qualified parking space, and merges and summarizes them by parking space to generate the reuse remaining time quantity; The time matching mapping submodule retrieves the vehicle request records in the concentrated distribution section based on the remaining reuse time, compares the remaining available time period of the parking space with the vehicle request time requirement, determines whether the usage interval overlap condition is met, performs time mapping between each parking space and the qualified request, and establishes a corresponding table for the prediction of reusable parking spaces.

6. The big data-based smart city parking resource prediction and optimization decision-making platform according to claim 1 is characterized in that: The parking space allocation decision module includes: The tolerance matching judgment submodule calculates the time difference based on the reusable duration and vehicle request duration data of each docking relationship in the reusable parking space prediction corresponding table, determines whether the difference is within the time tolerance range, screens the matching relationships between the qualified parking spaces and requests, and obtains a set of candidate allocation relationships; The candidate ratio calculation submodule classifies and counts the candidate allocation relationship set by parking section. In each parking section, the number of candidate allocation spaces and the corresponding number of vehicle requests are extracted. Based on the basic allocation standard set for parking supply adequacy in municipal planning, the ratio of the two is compared to see if it exceeds the matching benchmark threshold. Supply-demand ratio data for each parking section is then obtained to generate an allocation supply-demand ratio sequence. The priority push marking submodule selects parking sections with supply-demand ratio values ​​higher than the set benchmark ratio according to the allocation supply-demand ratio sequence, sets the push priority flag, organizes the pushable objects and their allocation numbers in combination with the corresponding candidate allocation relationship information, and establishes a parking space allocation table to be pushed.

7. The big data-based smart city parking resource prediction and optimization decision-making platform according to claim 1 is characterized in that: The system further comprises: The trend feedback optimization module extracts the response rate, usage completion rate, and response delay time of each allocated parking space based on the parking space allocation table to be pushed, determines the adaptability of each parking section allocation mode, and combines the multi-section solution structure based on the parking density trend in the subsequent time period to generate a parking resource optimization management plan; The parking resource optimization management plan includes statistical records of parking space allocation response rates in each section, parking space utilization completion rate analysis results, allocation response delay evaluation results, and multi-section allocation combination structure.

8. The big data-based smart city parking resource prediction and optimization decision-making platform according to claim 7 is characterized in that: The trend feedback optimization module includes: The response data extraction submodule extracts the response rate data, usage completion rate data, and response delay time data of each parking space based on the parking space allocation table to be pushed, obtains parking space response records, parking payment confirmation information, and user operation timestamps, performs data cross-validation, and integrates them into a time series format to obtain a deployment response indicator sequence; The deployment adaptation determination submodule normalizes the response rate, usage completion rate, and response delay time of each parking section based on the deployment response indicator sequence, calculates the difference between the response rate and the response time limit standard, calculates the average matching degree according to the indicator dimension, determines the adaptability of the deployment mode of each section, and obtains the deployment adaptation analysis results; The optimization combination generation submodule divides the parking sections into groups according to the allocation fitness analysis results and the parking density trend information in the subsequent time period, sets the combination structure matching, constructs the parking resource allocation logic relationship, and establishes the parking resource optimization management plan.

Images