A passenger flow heat support advertisement putting area dynamic selection resource allocation method

Abstract

The application discloses a kind of passenger flow thermodynamic support's advertisement delivery area dynamic selection resource allocation method, collect the multi-source passenger flow thermodynamic data in target city range;The target city is divided into several candidate advertisement delivery areas;Combining the advertisement site resource attribute, passenger flow attribute and commercial attribute of each candidate area, the comprehensive value score of each candidate area is calculated Advertisement delivery;According to advertisement delivery budget constraint and delivery target;Establish resource allocation optimization model;Real-time monitoring passenger flow thermodynamic change and advertisement delivery effect.The application has beneficial effects: fusion multi-source passenger flow thermodynamic data, from density, time length, frequency three dimensions quantification area thermodynamic, replace traditional manual experience point selection;Establish mathematical optimization model, solve the total value maximization resource allocation scheme under budget constraint, compared with fixed proportion allocation;Support minute-level passenger flow monitoring and incremental dynamic adjustment, can quickly respond to morning and evening peak, holiday, sudden event and other passenger flow change scenarios.

Description

Technical Field

[0001] This invention relates to the field of advertising placement technology, and in particular to a method for dynamically selecting and allocating resources for advertising placement areas based on passenger flow heat maps. Background Technology

[0002] With the rapid development of outdoor advertising and digital media, the requirements for the precision of advertising placement are constantly increasing. Traditional advertising placement mainly relies on manual experience to select placement areas and allocate resources according to a fixed ratio, which has the following drawbacks: First, it cannot capture changes in customer flow in real time, and the selection of placement areas lags behind customer flow; second, resource allocation lacks quantitative basis, which easily leads to insufficient resources in high-customer-flow areas and wasted resources in low-customer-flow areas; third, it is difficult to make dynamic adjustments based on the effect of the campaign, resulting in a low overall ROI.

[0003] Existing technologies include some advertising methods based on passenger flow data, but most simply use the number of passengers as a single metric, failing to comprehensively consider multi-dimensional factors such as passenger dwell time, visit frequency, and customer group matching. Furthermore, resource allocation often employs a fixed-ratio method, lacking mathematical optimization models and thus failing to maximize value within budget constraints. In addition, existing methods lack dynamic adjustment capabilities, making it difficult to cope with scenarios involving drastic fluctuations in passenger flow, such as peak hours and holidays. Therefore, it is necessary to propose a dynamic resource allocation method for advertising placement areas based on passenger flow heatmaps to address the aforementioned technical issues. Summary of the Invention

[0004] This invention provides a method for dynamically selecting and allocating resources for advertising placement based on passenger flow heat data. It calculates the regional heat index by using multi-source passenger flow heat data, constructs a comprehensive value scoring system by combining multi-dimensional attributes, establishes an optimization model to achieve optimal resource allocation, and supports real-time dynamic adjustment, significantly improving the accuracy and return on investment of advertising placement.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A method for dynamically selecting and allocating advertising resources based on passenger flow heatmaps includes the following steps: S1. Collect multi-source passenger flow heat map data within the target city area, perform data cleaning and standardization processing, and obtain a standardized passenger flow dataset; S2. Divide the target city into several candidate advertising areas, and calculate the real-time passenger flow heat index of each candidate area based on the standardized passenger flow dataset; S3. Calculate the comprehensive advertising value score for each candidate region by combining the advertising space resource attributes, customer flow attributes, and commercial attributes of each candidate region. S4. Based on the advertising budget constraints and advertising objectives, dynamically filter regions according to the comprehensive value score to generate an initial combination of advertising regions; S5. Establish a resource allocation optimization model to optimally allocate advertising resources within the initial advertising area combination and generate an initial resource allocation plan. S6. Monitor changes in passenger flow and advertising effectiveness in real time. When dynamic adjustment conditions are triggered, update the combination of advertising areas and resource allocation plan.

[0006] The multi-source passenger flow heat map data mentioned in step S1 includes operator signaling data, mobile device location data, public transportation card swiping data, commercial venue access control data, and video passenger flow statistics; the data cleaning includes outlier removal, duplicate data deduplication, and missing value imputation; and the standardization processing includes data normalization and spatiotemporal dimension alignment.

[0007] The formula for calculating the real-time passenger flow heat index in step S2 is as follows: ; in Let i be the real-time passenger flow heat index of the i-th candidate region. This represents the real-time passenger flow density per unit area in the region. This represents the average length of stay for passengers in this area. This represents the frequency of visitor traffic to the area within the statistical period. These are the weighting coefficients corresponding to passenger flow density, dwell time, and visit frequency, respectively, and satisfy the following conditions: .

[0008] The candidate advertising placement area mentioned in step S2 adopts either the grid division method or the administrative region division method: the grid division method divides the target city into square grids with a side length of 50-500 meters; the administrative region division method uses streets, business districts or communities as the basic division units.

[0009] The formula for calculating the overall value score of advertising placement in step S3 is as follows: ; in Score the overall advertising value for the i-th candidate region. This is the real-time passenger flow heat index for the area. This represents the abundance coefficient of advertising space resources in the region. This represents the matching coefficient between the area's passenger flow and the target customer group. This represents the commercial consumption level coefficient for the region. These are the weights of the corresponding indicators, and they satisfy... .

[0010] The dynamic regional screening in step S4 specifically includes: setting a comprehensive value score threshold, screening out candidate regions with scores higher than the threshold; then prioritizing regions according to the target type, and selecting the top N regions to form an initial deployment region combination, where N is determined by the total budget and the average deployment cost of the region.

[0011] The resource allocation optimization model described in step S5 aims to maximize the total deployment value. Its objective function and constraints are as follows: ; in The total value of advertising spending, The value of a unit resource allocation in the i-th region. The amount of advertising resources allocated to the i-th region. Let B be the unit resource allocation cost for the i-th region, B be the total allocation budget, and n be the total number of regions in the initial allocation region combination.

[0012] The dynamic adjustment conditions mentioned in step S6 include: the fluctuation range of the passenger flow heat index in any area exceeds the preset threshold, the conversion rate of advertising effectiveness is lower than the preset lower limit, the total budget changes, and new high-value candidate areas appear.

[0013] The weighting coefficient Dynamic adjustments based on industry type for advertising placement: Increased advertising for fast-moving consumer goods (FMCG) products. The weight of luxury goods advertising has increased. and The weight of community service advertisements has increased. The weight.

[0014] A method for dynamically selecting and allocating advertising placement areas based on passenger flow heatmaps also includes a historical data correction step: retrieving historical passenger flow heatmap data for the same period over the past 30 days, calculating the historical average heatmap value, and smoothing and correcting the real-time passenger flow heatmap index using the following formula: ; in This is the corrected thermal index. This represents the average heat value for the same historical period. These are the weighting coefficients for real-time data.

[0015] In step S6, the real-time monitoring frequency is once every 15 minutes, and it is increased to once every 5 minutes during peak passenger flow periods; the dynamic adjustment adopts an incremental adjustment method, and the resource change ratio of a single adjustment does not exceed 20% of the total resources.

[0016] A method for dynamically selecting and allocating resources for advertising placement based on customer flow heatmaps also includes a closed-loop feedback process: collecting data on ad exposure, clicks, and conversions in each placement area, calculating the actual ROI, and adjusting the overall value score weight of each area based on the ROI results.

[0017] A dynamic resource allocation method for advertising placement based on customer flow heatmaps is proposed. When there are multiple advertisers, a multi-target resource allocation mechanism is adopted, which calculates the target customer matching degree for each advertiser separately, allocates advertising space resources in the same area according to the budget ratio, and avoids resource conflicts between advertisers of the same category.

[0018] A dynamic resource allocation method for advertising placement based on passenger flow heat map is proposed. This method is implemented on a cloud computing platform. Passenger flow heat map data is accessed in real time through a data interface, and the resource allocation plan is distributed to various advertising terminal devices through an API interface. It supports unified scheduling of various advertising carriers such as outdoor screens, community light boxes, and subway media.

[0019] Compared with the prior art, the present invention has the following advantages: 1. Integrate multi-source passenger flow heat data to quantify regional heat from three dimensions: density, duration, and frequency, replacing traditional manual experience-based site selection, improving the accuracy of regional selection by more than 30%; establish a mathematical optimization model to solve the resource allocation scheme that maximizes total value under budget constraints, improving the ROI of deployment by 25%-40% compared to fixed-ratio allocation.

[0020] 2. Supports minute-level passenger flow monitoring and incremental dynamic adjustment, enabling rapid response to passenger flow changes during peak hours, holidays, and emergencies; comprehensively considering four dimensions—passenger flow heat, advertising resources, customer group matching, and commercial level—the scoring system can flexibly configure weights according to industry characteristics to adapt to the advertising needs of all categories. Attached Figure Description

[0021] Figure 1 This is a flowchart of the method for dynamically selecting and allocating resources for advertising placement areas according to the present invention; Figure 2 This is a comparison chart showing the improvement trend of advertising performance in different scenarios according to the present invention; Figure 3 This is a comparison chart showing the configuration trends of core parameters for different deployment scenarios of the present invention. Detailed Implementation

[0022] To enable those skilled in the art to better understand and implement the technical solutions of the present invention, the present invention will be further described below in conjunction with specific embodiments and accompanying drawings. However, the embodiments described are not intended to limit the present invention.

[0023] like Figures 1 to 3As shown, a method for dynamically selecting and allocating advertising resources based on passenger flow heatmaps includes the following steps: S1. Collect multi-source passenger flow heat map data within the target city area, perform data cleaning and standardization processing, and obtain a standardized passenger flow dataset; S2. Divide the target city into several candidate advertising areas, and calculate the real-time passenger flow heat index of each candidate area based on the standardized passenger flow dataset; S3. Calculate the comprehensive advertising value score for each candidate region by combining the advertising space resource attributes, customer flow attributes, and commercial attributes of each candidate region. S4. Based on the advertising budget constraints and advertising objectives, dynamically filter regions according to the comprehensive value score to generate an initial combination of advertising regions; S5. Establish a resource allocation optimization model to optimally allocate advertising resources within the initial advertising area combination and generate an initial resource allocation plan. S6. Monitor changes in passenger flow and advertising effectiveness in real time. When dynamic adjustment conditions are triggered, update the combination of advertising areas and resource allocation plan.

[0024] The multi-source passenger flow heat map data mentioned in step S1 includes operator signaling data, mobile device location data, public transportation card swiping data, commercial venue access control data, and video passenger flow statistics; the data cleaning includes outlier removal, duplicate data deduplication, and missing value imputation; and the standardization processing includes data normalization and spatiotemporal dimension alignment.

[0025] The formula for calculating the real-time passenger flow heat index in step S2 is as follows: ; in Let i be the real-time passenger flow heat index of the i-th candidate region. This represents the real-time passenger flow density per unit area in the region. This represents the average length of stay for passengers in this area. This represents the frequency of visitor traffic to the area within the statistical period. These are the weighting coefficients corresponding to passenger flow density, dwell time, and visit frequency, respectively, and satisfy the following conditions: .

[0026] The candidate advertising placement area mentioned in step S2 adopts either the grid division method or the administrative region division method: the grid division method divides the target city into square grids with a side length of 50-500 meters; the administrative region division method uses streets, business districts or communities as the basic division units.

[0027] The formula for calculating the overall value score of advertising placement in step S3 is as follows: ; in Score the overall advertising value for the i-th candidate region. This is the real-time passenger flow heat index for the area. This represents the abundance coefficient of advertising space resources in the region. This represents the matching coefficient between the area's passenger flow and the target customer group. This represents the commercial consumption level coefficient for the region. These are the weights of the corresponding indicators, and they satisfy... .

[0028] The dynamic regional screening in step S4 specifically includes: setting a comprehensive value score threshold, screening out candidate regions with scores higher than the threshold; then prioritizing regions according to the target type, and selecting the top N regions to form an initial deployment region combination, where N is determined by the total budget and the average deployment cost of the region.

[0029] The resource allocation optimization model described in step S5 aims to maximize the total deployment value. Its objective function and constraints are as follows: ; in The total value of advertising spending, The value of a unit resource allocation in the i-th region. The amount of advertising resources allocated to the i-th region. Let B be the unit resource allocation cost for the i-th region, B be the total allocation budget, and n be the total number of regions in the initial allocation region combination.

[0030] The dynamic adjustment conditions mentioned in step S6 include: the fluctuation range of the passenger flow heat index in any area exceeds the preset threshold, the conversion rate of advertising effectiveness is lower than the preset lower limit, the total budget changes, and new high-value candidate areas appear.

[0031] The weighting coefficient Dynamic adjustments based on industry type for advertising placement: Increased advertising for fast-moving consumer goods (FMCG) products. The weight of luxury goods advertising has increased. and The weight of community service advertisements has increased. The weight.

[0032] A method for dynamically selecting and allocating advertising placement areas based on passenger flow heatmaps also includes a historical data correction step: retrieving historical passenger flow heatmap data for the same period over the past 30 days, calculating the historical average heatmap value, and smoothing and correcting the real-time passenger flow heatmap index using the following formula: ; in This is the corrected thermal index. This represents the average heat value for the same historical period. These are the weighting coefficients for real-time data.

[0033] In step S6, the real-time monitoring frequency is once every 15 minutes, and it is increased to once every 5 minutes during peak passenger flow periods; the dynamic adjustment adopts an incremental adjustment method, and the resource change ratio of a single adjustment does not exceed 20% of the total resources.

[0034] A method for dynamically selecting and allocating resources for advertising placement based on customer flow heatmaps also includes a closed-loop feedback process: collecting data on ad exposure, clicks, and conversions in each placement area, calculating the actual ROI, and adjusting the overall value score weight of each area based on the ROI results.

[0035] A dynamic resource allocation method for advertising placement based on customer flow heatmaps is proposed. When there are multiple advertisers, a multi-target resource allocation mechanism is adopted, which calculates the target customer matching degree for each advertiser separately, allocates advertising space resources in the same area according to the budget ratio, and avoids resource conflicts between advertisers of the same category.

[0036] A dynamic resource allocation method for advertising placement based on passenger flow heat map is proposed. This method is implemented on a cloud computing platform. Passenger flow heat map data is accessed in real time through a data interface, and the resource allocation plan is distributed to various advertising terminal devices through an API interface. It supports unified scheduling of various advertising carriers such as outdoor screens, community light boxes, and subway media.

[0037] Compared with the prior art, the present invention has the following advantages: 1. Integrate multi-source passenger flow heat data to quantify regional heat from three dimensions: density, duration, and frequency, replacing traditional manual experience-based site selection, improving the accuracy of regional selection by more than 30%; establish a mathematical optimization model to solve the resource allocation scheme that maximizes total value under budget constraints, improving the ROI of deployment by 25%-40% compared to fixed-ratio allocation.

[0038] 2. Supports minute-level passenger flow monitoring and incremental dynamic adjustment, enabling rapid response to passenger flow changes during peak hours, holidays, and emergencies; comprehensively considering four dimensions—passenger flow heat, advertising resources, customer group matching, and commercial level—the scoring system can flexibly configure weights according to industry characteristics to adapt to the advertising needs of all categories.

[0039] Example 1: This embodiment takes the outdoor large screen advertising in the core business district of a certain city as an example to explain in detail the specific implementation process of this method.

[0040] Step S1: Data Acquisition and Preprocessing. Access operator signaling data, subway card swipe data, and commercial area video passenger flow data. Clean the raw data: remove outlier data with positioning errors greater than 100 meters, eliminate duplicate records, and use linear interpolation to complete missing data. Unify all data to the WGS84 coordinate system and standardize the time granularity to 15 minutes to generate a standardized passenger flow dataset.

[0041] Step S2: Area Division and Heat Index Calculation. A 200m × 200m grid method was used to divide the target business district into 86 candidate areas. The real-time passenger flow heat index for each area was calculated using the formula: ; In this embodiment, the weights for passenger flow density (α=0.5), dwell time (β=0.3), and visit frequency (γ=0.2) are calculated. The core pedestrian street area has the highest heat index at 92.5, while the peripheral area has the lowest at 23.1.

[0042] Step S3: Calculate the overall value score by collecting resource attributes such as the number, size, and visibility angle of ad placements in each region, and combining them with customer age structure and spending power data. ; The weightings are as follows: heat index 0.4, resource abundance 0.3, customer group matching degree 0.2, and consumption level 0.1.

[0043] Step S4: Initial Targeting Area Selection. A scoring threshold of 60 points was set, resulting in 32 candidate areas. These areas were then sorted from highest to lowest score, and based on a total budget of 1 million yuan, the top 18 areas were selected to form the initial targeting portfolio.

[0044] Step S5: Resource Optimization and Allocation. An optimization model is established to maximize the total investment value and determine the resource allocation for each area. Using the simplex method, the following allocations are determined: 45% of the budget is allocated to the three core pedestrian areas, 35% to the five secondary core areas, and 20% to the remaining ten areas.

[0045] Step S6: Dynamically adjust the heat fluctuation threshold to ±20%. During the evening peak hours, if the heat of passenger flow in the subway exit area increases by 35%, the adjustment mechanism is triggered, and 10% of the incremental resources are allocated from the edge areas to the subway exit area, so as to realize the dynamic flow of resources with passenger flow.

[0046] The campaign lasted 7 days, reaching a total of 12.8 million impressions, with an actual conversion rate of 3.2%. Compared to traditional fixed-area campaigns, the ROI increased by 37.5%.

[0047] Example 2: This embodiment takes the placement of light boxes and passageway advertisements in subway stations throughout a certain city as an example to focus on verifying the minute-level dynamic adjustment capability driven by public transportation passenger flow data.

[0048] Step S1: Data Acquisition and Preprocessing. This involves accessing three data sources: subway AFC card swiping data, station video passenger flow statistics, and operator subway station domain signaling data. Data Cleaning: Abnormal card swiping records (the same card repeatedly entering and exiting the station within 10 minutes) are removed, and passenger flow data for periods of equipment failure is supplemented (using comparative interpolation between adjacent stations). Standardization: Spatially, a standardized radiation area is defined centered on the subway station entrance / exit; temporally, a standardized granularity of 5 minutes is used to generate a standardized subway station-level passenger flow dataset.

[0049] Step S2: Area Division and Heat Index Calculation. The "station radial grid" division method is used. Each subway station entrance / exit is centered on a radial grid with a side length of 100 meters, serving as candidate deployment areas. A total of 127 candidate areas are generated citywide. The real-time passenger flow heat index is calculated using the formula: This embodiment is for a commuting scenario, so the weight of passenger flow density is increased. Dwell time weight Visit frequency weight During the morning rush hour, the heat index of core transfer stations reached a peak of 96.8, while that of suburban stations was the lowest at 18.3.

[0050] Step S3: Calculate the overall value score by collecting resource attributes such as the number, size, and viewing distance of ad slots at each site, combining these with the age and occupational tags of commuter customers to calculate customer matching degree, and then calculating the overall value score. The weighting is as follows: heat index 0.5, resource abundance 0.3, customer matching degree 0.15, and consumption level 0.05, which is suitable for the needs of commuting-related advertising.

[0051] Step S4: Initial Deployment Area Selection. A scoring threshold of 55 points was set, resulting in 72 candidate areas. Based on a monthly deployment budget of 800,000 yuan, the top 45 station areas were selected according to their scores to form the initial deployment combination, covering over 90% of the city's subway passenger flow.

[0052] Step S5: Resource Optimization and Allocation. An optimization model for resource allocation is established, aiming to maximize total exposure value and solve for the allocation scheme. Through linear programming, the solution is: 55% of the budget is allocated to the 10 core transfer stations, 30% to the 20 backbone stations, and 15% to the remaining 15 stations.

[0053] Step S6: Dynamically adjust the heat flow fluctuation threshold to ±25%, and monitor the frequency every 5 minutes during morning peak (7:00-9:00) and evening peak (17:00-19:00), and every 15 minutes during off-peak hours. One morning during peak hours, a sudden rainfall occurred, and the heat flow at three stations near office buildings increased by 40%, triggering the incremental adjustment mechanism. 12% of advertising resources were transferred from off-peak stations in the suburbs to these three stations. This single resource adjustment of 12% complies with the constraint of not exceeding 20%.

[0054] Example 3: This embodiment takes the placement of lightbox media in urban residential communities as an example, involving advertisers from three different categories: fast-moving consumer goods, children's education, and fitness clubs, to verify the multi-objective resource allocation and category conflict avoidance mechanism.

[0055] Step S1: Data Collection and Preprocessing. This involves accessing community access control data, location data from nearby mobile operators, and customer flow data from community shopping malls. Data Cleaning: Temporary visitor data (excluding non-residents) is removed, duplicate access control records are deduplicated, and missing daily data is supplemented using weekly average data. Standardization: Using the community as the basic spatial unit and a uniform time granularity of 1 hour, a community-level customer flow dataset is generated.

[0056] Step S2: Regional Division and Heat Index Calculation. The administrative division method was used, with each individual community as a basic candidate distribution area, encompassing a total of 216 residential communities. The real-time passenger flow heat index for each community was calculated: ; This scenario pertains to community services, and therefore prioritizes increasing visit frequency. The focus is on covering permanent residents who visit frequently.

[0057] Step S3: Calculate customer segment matching for each of the three advertisers based on their overall value score: FMCG (Fast Moving Consumer Goods) matches residents of all ages; children's education matches families with children aged 6-12; and fitness clubs match young adults aged 18-45. Three separate overall value scoring systems are generated, all using the following formula: Where k represents the k-th advertiser. Let be the customer matching coefficient of the k-th advertiser in the i-th community, with the weight set according to the category differentiation.

[0058] Step S4: Select three advertisers by region with budgets of 300,000, 250,000, and 200,000 respectively, for a total budget of 750,000. Target regions are then selected based on their respective comprehensive value scores: 120 communities for fast-moving consumer goods (FMCG) and beverages, 86 communities for children's education, and 72 communities for fitness clubs.

[0059] Step S5: Multi-objective resource allocation adopts a multi-objective resource allocation mechanism, allocating ad space resources within the same community according to the budget ratio of each advertiser, while setting mutual exclusion rules for the same product category. A joint optimization model is established to maximize the individual placement value of the three advertisers under the total resource constraint, ultimately generating a joint placement plan with a community media resource reuse rate of 82%.

[0060] Step S6: Performance Feedback and Weight Adjustment. Collect weekly data on ad scans and in-store conversions for each community, and calculate the regional ROI for each advertiser. After two weeks of operation, adjust the scoring weights based on the ROI data: increase the customer matching weight to 0.35 for children's education products and increase the customer density weight to 0.45 for FMCG products, achieving closed-loop performance optimization.

[0061] Implementation results: The average accuracy of reaching the target customer groups of the three advertisers increased by 29%, the overall utilization rate of community advertising space increased by 45%, and the overall advertising cost decreased by 22% compared with independent advertising by a single advertiser.

[0062] Example 4: This embodiment takes the temporary promotional advertising placement in a core business district of a city during the National Day Golden Week as an example to verify the historical data correction mechanism and the ability to adjust short-term dynamic budgets.

[0063] Step S1: Data Acquisition and Preprocessing: Access historical National Day holiday passenger flow data, real-time mobile location data, and commercial area video passenger flow data. Data Cleaning: Remove extremely abnormal group passenger flow data and supplement data in equipment blind spots. Standardization Processing: Spatially align with commercial area grid coordinates, unify time granularity to 10 minutes, and retrieve historical passenger flow data for the same period during the past 3 National Day holidays.

[0064] Step S2: Heat Index Calculation and Historical Correction. A 200m × 200m grid is used, resulting in 92 candidate areas. First, the real-time passenger flow heat index is calculated, then smoothed and corrected using historical data. The correction formula is: ; Among them, real-time data weight Historical average heat value for the same period The data is taken from the average of the same period during the National Day holiday over the past three years to avoid misjudgments caused by sudden fluctuations in real-time passenger flow. After correction, the heat index of the core business district's central area remains stable in the range of 88-94.

[0065] At the same time, the weighting of the heat index is dynamically adjusted according to the industry attributes of fast-moving consumer goods promotions: , , It focuses on passenger flow density and visit frequency.

[0066] Step S3: Calculate the overall value score by combining temporary advertising space resources (outdoor screens, mobile display stands, light boxes), the matching degree of the target customer group for the promotion, and the consumption level of the business district. The weights are as follows: the revised heat index is 0.45, the temporary resource abundance is 0.25, the customer group matching degree is 0.2, and the consumption level is 0.1.

[0067] Step S4: Initial Targeting Area Screening. A scoring threshold of 65 points was set to select 48 candidate areas. Based on a temporary budget of 500,000 yuan for 7 days, the top 28 high-value areas were selected to form the initial targeting portfolio, covering the core pedestrian paths and main entrances / exits of the business district.

[0068] Step S5: Resource Optimization and Allocation. Establish a short-term campaign optimization model to maximize the total promotional reach value and solve for a daily resource allocation plan. The solution yields: allocate 65% of the budget to the peak traffic days before National Day (3 days in advance) and 35% to the following 4 days; allocate 52% of the total budget to the core pedestrian street area.

[0069] Step S6: Budget Change Triggers Adjustment Day 3: The brand adds 200,000 in budget, triggering the budget change adjustment conditions. The optimization model is rerun, adding 10 secondary core area placement points. At the same time, 70% of the incremental budget is allocated to the catering and entertainment areas with continuously increasing customer traffic, and 30% is supplemented to the core areas.

[0070] The total exposure of promotional advertisements during this Golden Week reached 9.8 million people, driving a 27% increase in store visits to shopping districts for promotional activities. Compared with fixed-location advertising in the same period of previous years, the ROI of the promotion increased by 34.8%.

[0071] Based on the above embodiments 1-4, the following lists 1 and tables 2 are derived: Table 1 compares the core performance indicators for different advertising scenarios;

[0072] Table 2 compares the core algorithms and runtime parameter configurations for different scenarios;

[0073] Based on Tables 1 and 2 above, the following conclusions can be drawn. The method is universally applicable across all scenarios and significantly improves performance. In four typical scenarios, the ROI of advertising campaigns has been steadily improved in the range of 30% to 38%, and the core business indicators (exposure rate, store visit rate, and cost control) have all achieved optimization of more than 20%. This verifies the dynamic site selection and resource allocation method driven by customer flow heat, and has universal value for mainstream advertising campaigns such as outdoor advertising, transportation media, community media, and temporary promotions.

[0074] The parameters can be flexibly adapted to different scenarios and objectives, and the differentiated configuration provides high precision. The method can match different targeting goals by adjusting the weight of the heat index: for commuter subway scenarios, the weight of passenger density is increased (α=0.6) to prioritize covering high-traffic groups; for community scenarios, the weight of visit frequency is increased (γ=0.35) to focus on reaching permanent residents; for holiday promotion scenarios, historical data is introduced for smoothing correction (δ=0.7) to reduce interference from sudden fluctuations in passenger flow. This dynamic adaptability of the parameters is the core reason why improved results can be achieved across different scenarios.

[0075] Minute-level dynamic adjustments can fully capture the value of passenger flow fluctuations. In subway scenarios, by using 5-minute-level monitoring and incremental resource adjustments during peak hours, the effective exposure rate during peak hours has been increased by 42%. This shows that real-time passenger flow monitoring combined with a dynamic resource scheduling mechanism can effectively capture the traffic dividends of drastic passenger flow fluctuations during morning and evening peak hours and holidays, and solve the structural problem of "insufficient resources during high passenger flow and wasted resources during low passenger flow" in traditional fixed deployment.

[0076] The multi-advertiser model can significantly improve the efficiency of media resource utilization. In the scenario of joint advertising by multiple advertisers in the community, the media resource utilization rate increases by 45% and the overall advertising cost decreases by 22%. This proves that multi-target resource allocation and the same category conflict avoidance mechanism can achieve efficient reuse of the same media resources by multiple advertisers. While ensuring the accuracy of each advertiser's advertising, it can significantly reduce the advertising cost per customer.

[0077] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application also intends to include these modifications and variations.

Claims

1. A method for dynamically selecting and allocating resources for advertising placement based on passenger flow heatmaps, characterized in that: Includes the following steps: S1. Collect multi-source passenger flow heat map data within the target city area, perform data cleaning and standardization processing, and obtain a standardized passenger flow dataset; S2. Divide the target city into several candidate advertising areas, and calculate the real-time passenger flow heat index of each candidate area based on the standardized passenger flow dataset; S3. Calculate the comprehensive advertising value score for each candidate region by combining the advertising space resource attributes, customer flow attributes, and commercial attributes of each candidate region. S4. Based on the advertising budget constraints and advertising objectives, dynamically filter regions according to the comprehensive value score to generate an initial combination of advertising regions; S5. Establish a resource allocation optimization model to optimally allocate advertising resources within the initial advertising area combination and generate an initial resource allocation plan. S6. Monitor changes in passenger flow and advertising effectiveness in real time. When dynamic adjustment conditions are triggered, update the combination of advertising areas and resource allocation plan.

2. The method for dynamically selecting and allocating resources for advertising placement based on passenger flow heat map according to claim 1, characterized in that: The multi-source passenger flow heat map data mentioned in step S1 includes operator signaling data, mobile device location data, public transportation card swiping data, commercial venue access control data, and video passenger flow statistics; the data cleaning includes outlier removal, duplicate data deduplication, and missing value imputation; and the standardization processing includes data normalization and spatiotemporal dimension alignment.

3. The method for dynamically selecting and allocating resources for advertising placement areas based on passenger flow heat map according to claim 2, characterized in that: The formula for calculating the real-time passenger flow heat index in step S2 is as follows: ； in Let i be the real-time passenger flow heat index of the i-th candidate region. This represents the real-time passenger flow density per unit area in the region. This represents the average length of stay for passengers in this area. This represents the frequency of visitor traffic to the area within the statistical period. These are the weighting coefficients corresponding to passenger flow density, dwell time, and visit frequency, respectively, and satisfy the following conditions: .

4. The method for dynamically selecting and allocating resources for advertising placement based on passenger flow heat map according to claim 3, characterized in that: The candidate advertising placement area mentioned in step S2 adopts either the grid division method or the administrative region division method: the grid division method divides the target city into square grids with a side length of 50-500 meters; the administrative region division method uses streets, business districts or communities as the basic division units.

5. The method for dynamically selecting and allocating resources for advertising placement based on passenger flow heat map according to claim 4, characterized in that: The formula for calculating the overall value score of advertising placement in step S3 is as follows: ； in Score the overall advertising value for the i-th candidate region. This is the real-time passenger flow heat index for the area. This represents the abundance coefficient of advertising space resources in the region. This represents the matching coefficient between the area's passenger flow and the target customer group. This represents the commercial consumption level coefficient for the region. These are the weights of the corresponding indicators, and they satisfy... .

6. The method for dynamically selecting and allocating resources for advertising placement areas based on passenger flow heat map according to claim 5, characterized in that: The dynamic regional screening in step S4 specifically includes: setting a comprehensive value score threshold, screening out candidate regions with scores higher than the threshold; then prioritizing regions according to the target type, and selecting the top N regions to form an initial deployment region combination, where N is determined by the total budget and the average deployment cost of the region.

7. The method for dynamically selecting and allocating resources for advertising placement areas based on passenger flow heat map according to claim 6, characterized in that: The resource allocation optimization model described in step S5 aims to maximize the total deployment value. Its objective function and constraints are as follows: ； in The total value of advertising spending, The value of a unit resource allocation in the i-th region. The amount of advertising resources allocated to the i-th region. Let B be the unit resource allocation cost for the i-th region, B be the total allocation budget, and n be the total number of regions in the initial allocation region combination.

8. The method for dynamically selecting and allocating resources for advertising placement areas based on passenger flow heat map according to claim 7, characterized in that: The dynamic adjustment conditions mentioned in step S6 include: the fluctuation range of the passenger flow heat index in any area exceeds the preset threshold, the conversion rate of advertising effectiveness is lower than the preset lower limit, the total budget changes, and new high-value candidate areas appear.

9. The method for dynamically selecting and allocating resources for advertising placement based on passenger flow heat map according to claim 8, characterized in that: The weighting coefficient Dynamic adjustments based on industry type for advertising placement: Increased advertising for fast-moving consumer goods (FMCG) products. The weight of luxury goods advertising has increased. and The weight of community service advertisements has increased. The weight.

10. The method for dynamically selecting and allocating resources for advertising placement areas based on passenger flow heat map according to claim 9, characterized in that: It also includes a historical data correction step: retrieving historical passenger flow heat map data for the same period in the past 30 days, calculating the historical average heat map value, and smoothing and correcting the real-time passenger flow heat map index. The correction formula is: ； in This is the corrected thermal index. This represents the average heat value for the same historical period. The weighting coefficients are based on real-time data. In step S6, the real-time monitoring frequency is once every 15 minutes, increasing to once every 5 minutes during peak hours. Dynamic adjustments are made using an incremental adjustment method, with the resource change ratio for a single adjustment not exceeding 20% ​​of the total resources. The method also includes a closed-loop feedback step: collecting ad exposure, click-through rate, and conversion rate data for each advertising area, calculating the actual ROI, and reversing the overall value score weight for each area based on the ROI results. When multiple advertisers exist, a multi-target resource allocation mechanism is adopted, calculating the target customer group matching degree separately for each advertiser, allocating ad space resources within the same area according to the budget ratio, and avoiding resource conflicts between advertisers of the same category. This method is implemented based on a cloud computing platform, with real-time access to passenger flow heatmap data via a data interface, and resource allocation schemes distributed to various advertising terminal devices via API interfaces, supporting unified scheduling of various advertising carriers such as outdoor screens, community light boxes, and subway media.

Images