Traffic jam management optimization algorithm based on regional population scale and economic structure

Abstract

The invention relates to the technical field of traffic control optimization, and discloses a traffic jam control optimization algorithm based on a regional population scale and an economic structure. The algorithm comprises the following steps: firstly, obtaining population distribution data, economic industry data and real-time traffic flow original data of a target area, and generating an area traffic characteristic representative data set through multi-dimensional clustering processing; generating a traffic regulation and control allocation instruction set based on the data set; then constructing a dynamic traffic state hierarchical optimization model containing a top layer road network stability model and a bottom layer resource cost model; and finally, determining a cooperative constraint condition of the two models, inputting an instruction set into the models, performing cross-layer iteration updating solution according to the constraint condition, and outputting an optimal traffic control strategy configuration scheme. According to the algorithm, multi-dimensional data are fused, collaborative optimization of road network stability and resource cost is realized, the accuracy and adaptability of traffic jam treatment are improved, and regional traffic jam is effectively relieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of traffic management optimization, in particular to a traffic congestion management optimization algorithm based on regional population scale and economic structure. BACKGROUND

[0002] With the acceleration of urbanization, the scale of population flow in the region is expanding, and the economic industrial structure is continuously adjusted, and the pressure on the transportation system is increasing. At present, most traffic congestion management solutions still rely on single-dimensional data support, such as only using real-time traffic flow data to develop control strategies, without fully considering the deep influence of population distribution and economic industrial activities on traffic demand. In terms of population distribution, there are significant differences in residential population density and commuting population scale in different regions. The traffic flow distribution during morning and evening peak hours is often highly correlated with population gathering areas. If this factor is ignored, it is easy to cause the control measures to be out of touch with the actual traffic demand. From the perspective of economic industry, enterprises of different types of industries have obvious characteristics in spatial distribution, such as manufacturing enterprises are mostly concentrated in industrial parks, and service enterprises are densely distributed in urban central business districts. The traffic flow in these industrial gathering areas has specific rules. The freight vehicles in manufacturing areas come and go frequently, and the passenger flow in service areas has a long peak duration. The existing management means lack deep integration of economic industry data, making it difficult to accurately identify the causes of traffic congestion caused by industrial activities, resulting in insufficient pertinence of the developed management strategies. Current traffic optimization models mostly use single-level design, focusing on either the overall stability of the road network or only on the control of resource costs, without achieving the coordinated optimization of both. In actual application, if only the stability of the road network is pursued and excessive traffic resources are invested, it will cause resource waste; if only the cost is controlled and the level of resource allocation is reduced, it will affect the efficiency of the road network. In addition, the existing models lack a dynamic iteration mechanism in the solving process, and cannot adjust the optimization direction in time according to the changes in real-time traffic state, resulting in poor adaptability of the output management strategies and difficulty in coping with complex and variable traffic scenarios. The existence of these problems makes it difficult to effectively improve the current traffic congestion management effect, and cannot meet the needs of sustainable development of regional traffic. SUMMARY

[0003] The purpose of the present application is to provide a traffic congestion management optimization algorithm based on regional population scale and economic structure to solve the problems raised in the background.

[0004] To achieve the above purpose, the present application provides a traffic congestion management optimization algorithm based on regional population scale and economic structure, which comprises: Acquire population distribution data, economic and industrial data, and real-time traffic flow raw data of the target area, and perform multi-dimensional clustering processing on the population distribution data, economic and industrial data, and real-time traffic flow raw data to generate a representative dataset of regional traffic characteristics. Based on the representative dataset of regional traffic characteristics, a set of traffic control and allocation instructions is generated. A dynamic traffic state hierarchical optimization model is constructed based on the traffic control and allocation instruction set. The dynamic traffic state hierarchical optimization model includes a top-level road network stability model and a bottom-level resource cost model. The collaborative constraints of the top-level road network stability model and the bottom-level resource cost model are determined. The traffic control and allocation instruction set is input into the top-level road network stability model and the bottom-level resource cost model. Cross-level iterative update is performed based on the collaborative constraints to solve the problem and output the optimal traffic management strategy configuration scheme.

[0005] Preferably, generating a traffic control and allocation instruction set based on the regional traffic feature representative dataset includes: The representative dataset of regional traffic characteristics is subjected to dynamic gain Kalman filter decomposition to separate the high-frequency fluctuation component of traffic flow from the low-frequency trend component of traffic demand. A hybrid management system is selected, consisting of road capacity adjustment and traffic guidance strategies. The low-frequency trend component of traffic demand is used as the road network load benchmark value, and the high-frequency fluctuation component of traffic flow is dynamically allocated within the hybrid management system to generate a traffic control and allocation instruction set. Preferably, the step of performing multi-dimensional clustering processing on the population distribution data, economic industry data, and real-time traffic flow raw data to generate a representative dataset of regional traffic features includes: The population distribution data, economic industry data, and real-time traffic flow raw data are normalized to generate a standardized set of traffic parameters. The silhouette coefficient method is introduced to calculate the multidimensional clustering quality assessment index; Determine the silhouette coefficients for different numbers of clusters and construct silhouette coefficient variation curves. When the silhouette coefficient variation curves reach a preset stable threshold, output the optimal number of clusters. Based on the optimal number of clusters, initial cluster centroids are selected from the standardized traffic parameter set. The weighted Euclidean distances from all standardized traffic parameters to the initial cluster centroids are calculated. Each standardized traffic parameter is assigned to the cluster corresponding to the minimum weighted Euclidean distance. The cluster centroids are iteratively updated until the centroid offset is lower than the convergence threshold, so as to generate a representative dataset of regional traffic features.

[0006] Preferably, the step of performing dynamic gain Kalman filter decomposition on the representative dataset of regional traffic features includes: The regional traffic feature representative dataset is input into a variable gain Kalman filter for state prediction, and a mathematical model of regional traffic state evolution is established. Configure the time update equation and state correction equation of the mathematical model for the evolution of traffic state in the region; Based on historical traffic conditions, the current traffic flow observation is predicted. The current traffic flow observation is then input into the state correction equation for residual calibration, and the calibrated traffic state estimate is output. The calibrated traffic state estimate is input into the time update equation to predict the prior estimate of traffic state for the next period, so as to separate the high-frequency fluctuation component of traffic flow from the low-frequency trend component of traffic demand.

[0007] Preferably, the step of dynamically allocating the high-frequency fluctuation components of traffic flow within the hybrid management system to generate a traffic control allocation instruction set includes: The high-frequency fluctuation component of the traffic flow is used as the control quantity input of the hybrid governance system into the dynamic routing distributor. Establish a mathematical model for traffic control and allocation based on Kalman prediction; Configure the control quantity update equation and resource allocation equation of the traffic control and allocation mathematical model; Based on the traffic fluctuation components of the previous period, the current control demand value is predicted, and the current control demand value is input into the resource allocation equation for strategy correction to generate the current control strategy correction value. The current control strategy correction value is input into the control quantity update equation to predict the traffic control allocation quantity for the next cycle. Calculate the dynamic route allocation coefficients; Based on the dynamic routing allocation coefficient, a traffic control allocation rule is generated. According to the traffic control allocation rule, the predicted traffic control allocation amount for the next cycle is allocated to the road capacity adjustment unit and the traffic guidance strategy unit to generate a traffic control allocation instruction set.

[0008] Preferably, the top-level road network stability model takes minimizing the road network fluctuation index as its objective function; The underlying resource cost model includes four sub-models: initial investment cost of governance facilities, operation and maintenance cost, facility renewal cost, and traffic delay penalty cost. The underlying resource cost model takes maximizing the overall regional benefits as its objective function. The overall regional benefits consist of the traffic management revenue minus the initial investment cost, operation and maintenance cost, facility renewal cost, and traffic delay penalty cost of the management facilities, plus the environmental benefits of improved traffic efficiency.

[0009] Preferably, the collaborative constraints include: Road capacity adjustment constraints: The traffic capacity of road network nodes is between the preset minimum traffic threshold and the maximum carrying capacity threshold; Traffic guidance strategy constraints: The frequency of guidance information release meets the preset release cycle requirements.

[0010] Preferably, the collaborative constraint conditions further include: Hybrid governance system response constraints: The coordinated response time of the road capacity adjustment unit and the traffic guidance strategy unit is lower than the system response threshold; Road network stability constraints: The road network load change rate and the road network state dispersion shall not exceed the preset road network stability boundary.

[0011] Preferably, the collaborative constraint conditions further include: System supply and demand balance constraint: The total regional transportation supply equals the sum of the basic transportation demand and the regulation capacity of the mixed governance system; Economic structure constraints: The allocation ratio of transportation resources in each industrial cluster is consistent with the characteristics of the regional industrial economic structure.

[0012] Preferably, the step of inputting the traffic control and allocation instruction set into the top-level road network stability model and the bottom-level resource cost model, and performing cross-level iterative update and solution based on the collaborative constraints, includes: The traffic control and allocation instruction set is input into the top-level road network stability model to perform road network fluctuation analysis and generate a preliminary governance resource allocation plan. The preliminary governance resource allocation scheme is input into the underlying resource cost model for cost-benefit correction, generating a corrected resource cost model. The revised resource cost model integrates real-time traffic demand data to calculate the regulation requirements of the hybrid governance system. The regulation parameters corresponding to the preliminary governance resource allocation scheme are input into the road capacity adjustment unit model and the traffic guidance strategy unit model, respectively. With the goal of maximizing the comprehensive regional benefits, the model is solved in a two-way iterative manner under the cooperative constraints to output the optimal traffic governance strategy allocation scheme.

[0013] Compared with the prior art, the beneficial effects of the present invention are: By acquiring population distribution data, economic and industrial data, and real-time traffic flow data of the target area, and performing multi-dimensional clustering to generate a representative dataset of regional traffic characteristics, it is possible to comprehensively capture the key factors affecting traffic conditions. The introduction of population distribution data clearly presents the degree of population concentration and flow patterns in different areas, making traffic characteristic analysis more aligned with actual travel needs and avoiding control deviations caused by neglecting population factors. The integration of economic and industrial data accurately identifies traffic flow characteristics generated by industrial activities, such as the differences between freight and passenger traffic and the distribution of peak hours, providing comprehensive data support for the subsequent development of targeted governance strategies. The traffic control and allocation instruction set generated based on a representative dataset of regional traffic characteristics possesses greater scientific rigor and rationality. Compared to traditional methods that rely solely on traffic flow data to formulate instructions, this instruction set fully integrates information from population and economic dimensions, enabling precise allocation of traffic resources. This ensures that control measures directly address the causes of congestion and effectively reduce resource misallocation. Whether adjusting road traffic permissions or optimizing public transportation scheduling, this instruction set provides clear and practical guidance for specific operations. The constructed dynamic traffic state hierarchical optimization model comprises a top-level road network stability model and a bottom-level resource cost model, breaking through the limitations of traditional single-level models. The top-level road network stability model ensures the smooth operation of the entire traffic network from a macroscopic perspective, preventing overall network paralysis caused by the spread of localized congestion; the bottom-level resource cost model rationally controls the input of traffic resources from a microscopic perspective, avoiding excessive resource consumption. This hierarchical design allows for separate attention to road network stability and resource cost control. Furthermore, by determining the synergistic constraints between the two, it achieves coordinated optimization at both the macro and micro levels, ensuring efficient road network operation while also considering the rational use of resources, forming a virtuous cycle of mutual promotion and constraint. The application of cross-layer iterative update in the solution process enables dynamic adaptation to changes in traffic conditions. After inputting the traffic control and allocation instruction set into two hierarchical models, iterative updates are continuously performed based on collaborative constraints. This allows for timely correction of deviations in the model calculation process, ensuring that the output optimal traffic management strategy configuration is more consistent with real-time traffic scenarios. This dynamic adjustment mechanism effectively enhances the adaptability and flexibility of the solution, enabling it to cope with changes in traffic flow at different times and in different regions. Whether it's the large-scale commuter traffic during morning and evening rush hours or the cross-regional travel peaks during holidays, iterative solutions can output appropriate management strategies. Attached Figure Description

[0014] Figure 1 This is a schematic diagram illustrating the working principle of the traffic congestion management optimization algorithm based on regional population size and economic structure described in this invention. Figure 2 A flowchart for generating a traffic control and allocation instruction set; Figure 3 A flowchart for dynamic gain Kalman filter decomposition processing; Figure 4 Flowchart defining the objective functions for the top-level and bottom-level models; Figure 5 A flowchart for solving the optimal strategy through cross-layer iterative updates. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Please see Figure 1 This invention provides a traffic congestion management optimization algorithm based on regional population size and economic structure, the algorithm comprising: By integrating population distribution, economic industry, and real-time traffic flow data, a dynamic hierarchical optimization model for traffic conditions is constructed to achieve precise configuration of traffic management strategies. The algorithm first acquires raw data on population distribution, economic industry, and real-time traffic flow in the target area, and generates a representative dataset of regional traffic characteristics through multi-dimensional clustering. Based on this dataset, a traffic control and allocation instruction set is generated, and a dynamic hierarchical optimization model for traffic conditions, including a top-level road network stability model and a bottom-level resource cost model, is constructed. By determining collaborative constraints and performing cross-layer iterative updates, the optimal traffic management strategy configuration scheme is finally output.

[0017] Example 1: See Figure 2 Every day at 4:00 AM, the system automatically starts the data collection program, obtaining population density and age structure data for 78 communities within the region from the municipal population census database, including the number of permanent residents, employment distribution, and commuting mode ratios. Simultaneously, it extracts indicators such as the proportion of the tertiary sector, enterprise size distribution, and business activity intensity from the economic statistics platform. Traffic flow data comes from 216 roadside microwave detectors and floating car GPS tracks, updating vehicle speed, traffic volume, and occupancy rates for each road segment every 2 minutes.

[0018] During the data normalization process, the "number of residents per square kilometer" in the population data is mapped to the [0,1] interval, the "average daily freight frequency of enterprises" in the economic data is weighted and standardized according to industry type, and the traffic flow data is scaled using the maximum capacity at the minute level as the benchmark. A standardized traffic parameter set containing 387 dimensions is generated, with each dimension corresponding to a specific spatiotemporal unit (e.g., "Wednesday morning 8:00-8:15, South entrance of Financial Street").

[0019] The silhouette coefficient was calculated using the sliding window method. Initially, with K=3, the silhouette coefficient for the residential area, commercial area, and industrial park clusters was measured to be 0.62. When K=5, the coefficient increased to 0.78 after adding the transportation hub area and mixed-use area. Further increasing K to 7 resulted in only a 0.03 increase in the coefficient. The rate of change of the silhouette coefficient at K=5 was the first to fall below the stability threshold of 0.05, thus determining the optimal number of clusters as 5. The initial centroids were selected using the maximum spacing method. In the standardized parameter space, the residential area sample point with the highest commuting pressure was first marked as the first centroid. Then, the logistics park sample point with the greatest Euclidean distance was selected as the second centroid, and so on, determining five initial centroids. During distance calculation, commuting time data were assigned a weight of 1.5 times, and nighttime data a weight of 0.7 times, reflecting the timeliness requirements of traffic management.

[0020] During the iterative clustering process, the weighted distance between the sample points of a certain technology industrial park and the centroid of the industrial park was calculated as follows: population dimension difference 0.32 (weight 0.2), economic dimension difference 0.41 (weight 0.3), morning peak traffic flow difference 0.17 (weight 0.5), and comprehensive distance value 0.27. This value is less than the distance to the centroid of the commercial area (0.35), so it is classified into the industrial park cluster. After each iteration, the centroid position of each cluster is recalculated. The calculation stops when the centroid offset is less than 0.01 for three consecutive iterations, ultimately forming a representative dataset containing five types of traffic feature regions.

[0021] In the Kalman filter decomposition stage, a state-space model is established for the financial business district dataset. When establishing the state-space model for the financial business district dataset during the Kalman filter decomposition stage, it is necessary to define the system state variables. For example, key factors related to traffic within the financial business district, such as real-time traffic flow on main roads and average vehicle speed on road segments, are selected as components of the state vector. The state transition matrix is ​​determined, constructed based on the historical data showing the changes in traffic state over time, thus describing the evolution of the system state from the current moment to the next. Considering the influence of external factors on traffic, a control input matrix and corresponding control input vector are set. Due to the numerous uncertainties in the traffic system, process noise is introduced into the model, typically assumed to follow a specific distribution, such as Gaussian white noise. Regarding the observation equation, an observation matrix needs to be determined, constructed based on the mapping relationship between actually available observation data and system state variables. Observation noise is also considered to reflect errors and uncertainties in the observation process, thereby completely constructing a state-space model suitable for the financial business district dataset. In the time update equation, the traffic state at the same time of the previous working day is used as a benchmark, and the recent trend change rate is superimposed to predict the current state. On a Thursday morning at 8:00 AM, the predicted traffic flow on the main road was 1820 vehicles per hour, while the actual measured value was 1753 vehicles. The state correction equation, by adjusting the Kalman gain coefficient, calibrated the prediction model by scaling the residual of 67 vehicles to 0.7, resulting in a corrected estimated value of 1792 vehicles.

[0022] The low-frequency trend component is shown as follows: the baseline demand in the business district during the morning rush hour rose steadily from 1,420 vehicles at 7:30 to 2,130 vehicles at 9:15, forming a smooth upward curve. The high-frequency fluctuation component captures the 327-vehicle increase caused by the sudden subway malfunction at 8:45, and the 184-vehicle negative fluctuation caused by temporary traffic control at 8:03.

[0023] Based on the morning peak demand of 2,130 vehicles, the direction of the main road's tidal flow lanes is dynamically adjusted: four lanes are open for inbound traffic from 7:30 to 8:45, and after 8:45, four lanes are switched to outbound traffic. The traffic guidance strategy responds to high-frequency components; when a subway malfunction causes an increase of 327 vehicles, three alternative routes are immediately pushed through the navigation app to disperse the extra traffic flow to parallel side roads.

[0024] Through the above implementation process, the regional traffic characteristic dataset effectively distinguishes between routine commuting demand and sudden fluctuations, providing a precise control benchmark for the hybrid governance system. Road resource allocation focuses on predictable, regular demand, while dynamic guidance responds quickly to unforeseen traffic disturbances, forming a well-defined governance structure.

[0025] Example 2: See Figure 3 In the implementation of Example 2, an industrial logistics park in a coastal port city is used as an example. The traffic flow in this area exhibits significant characteristics of heavy truck transportation and is greatly affected by ship arrival times, with traffic fluctuations being sudden and highly correlated. Every day at 2:00 AM, the system obtains the ship berthing plans for the next 24 hours from the port authority's dispatch system, and simultaneously integrates container truck traffic data collected from 86 road detectors to form a representative dataset of regional traffic characteristics.

[0026] When the dynamic gain Kalman filter decomposition process is initiated, an industrial zone traffic state evolution model is first constructed, defining the traffic state as a three-dimensional vector: average vehicle speed on main roads, queue length at intersections, and the proportion of heavy vehicles. The decomposition process is as follows: First, based on historical data and analysis of traffic operation patterns in the industrial zone, a state transition matrix is ​​determined, reflecting the mutual influence and evolution of these three traffic state variables at different times. An observation matrix is ​​then determined to establish the connection between this three-dimensional traffic state vector and the actual observable data. During the filtering process, the state transition equation predicts the current state based on the state estimate from the previous time step. Then, combined with newly acquired observation data, the Kalman filter update equation is used to correct the predicted state, iterating continuously to achieve accurate estimation and decomposition of the industrial zone traffic state. In the time update equation, a baseline state is established based on data from the same time period of the previous day; for example, the predicted speed of the port access road is 28 km / h at 4:00 AM on Wednesday. At this point, a ship operation progress factor is introduced: when the "Dongfang" cargo ship completes unloading 3 hours ahead of schedule, the system automatically increases the expected value of freight demand and corrects the predicted vehicle speed for that period to 25km / h.

[0027] The state correction process relies on real-time detection data. At 3:50 AM on a Thursday, the radar detector at the port gate measured a vehicle speed of only 22 km / h, a 3 km / h deviation from the predicted value. The variable-gain Kalman filter activated its residual calibration mechanism: when the proportion of heavy vehicles exceeded 65%, it automatically increased the gain coefficient from 0.6 to 0.8, making the state estimate closer to the actual measured data. Calculations showed the vehicle speed estimate was adjusted to 22.4 km / h, while also separating the 4.2 km / h speed drop component caused by the temporary addition of 20 port-clearing trucks.

[0028] The dynamic adjustment of the gain coefficient follows specific rules. During peak ship arrival times (e.g., Tuesday morning 9:00-11:00), the system detected an increase in the number of arriving ships from 3 to 7 within three hours, with a 37% increase in freight demand volatility. At this time, the filter automatically increases the weight of the economic dimension from 0.3 to 0.5, enhancing the impact of industrial activity on state prediction. When freight transport was disrupted on Friday afternoon due to a typhoon warning, the system detected a sharp drop of 82% in traffic flow variance and immediately reduced the observation noise covariance matrix parameter by 40%, improving sensitivity to abnormal conditions. The dynamic adjustment of the gain coefficient follows specific rules based on the characteristics of traffic state changes and the difference between observed and predicted data. When traffic state changes are relatively stable and the deviation between observed and predicted data is small, the gain coefficient adjustment is small, making the filtering process rely more on model prediction results. However, when traffic state fluctuates significantly and the difference between observed and predicted data is significant, the gain coefficient adjustment increases accordingly, making the filtering process focus more on observed data. This adaptively balances the weights of model prediction and observed data in state estimation, thereby more accurately tracking the dynamic changes in traffic state.

[0029] The extraction of low-frequency trend components reveals a periodic pattern. Analysis of data from three consecutive weeks shows a clear diurnal rhythm in traffic demand at the logistics park: a peak in worker commuting occurs between 6:00 and 8:00 AM, a low point in logistics vehicle traffic occurs between 12:00 and 2:00 PM, and demand surges in the terminal operation area after 7:00 PM. Based on this, the system generates a baseline curve; for example, the demand in the terminal operation area stabilizes at 480 standard vehicle trips per hour at 8:00 PM every Friday.

[0030] High-frequency fluctuation components capture sudden disturbances. In one implementation, a temporary customs inspection at 10:15 AM on Wednesday caused a decrease in gate throughput efficiency. The system separated a 47-minute backlog of 122 vehicles from the total traffic flow. Simultaneously, it identified 86 positive fluctuations due to a sudden increase in orders from a certain company. This fluctuation exhibited a clear company clustering characteristic: 73% of the increased vehicles were heading towards the chemical storage area. Specifically, in traffic flow data, sudden disturbances may originate from unexpected situations such as traffic accidents, temporary traffic control, or large-scale events causing gatherings of people and vehicles. These sudden factors can cause drastic changes in traffic flow within a short period, forming high-frequency fluctuations. The high-frequency fluctuation components decomposed by Kalman filtering can keenly detect these abnormal changes, separating them from the overall traffic flow data. This provides crucial information for traffic management departments to promptly identify and respond to sudden traffic events, helping to quickly take appropriate measures to alleviate traffic congestion and other problems.

[0031] The component separation process adopts a multi-level filtering architecture. The first-level filter processes the impact of ship scheduling on a 15-minute scale and outputs the hourly freight demand baseline. The second level focuses on the intersection traffic fluctuations on a 5-minute scale and captures the platoon dispersion within a single signal cycle. The third level targets second-level sudden events, such as a truck breakdown that occurred at 14:03 on a Tuesday afternoon. The system identifies the 32-meter queue length increment caused by the event within 8 seconds.

[0032] In terms of data processing, the system establishes a dual-buffer system. The primary buffer stores the final component data after three rounds of filtering, while the backup buffer retains the original observation sequence. When data anomalies occur on Friday morning due to detector malfunction, the system automatically switches to the backup data source, using the average of the previous three days' data for the same period for status prediction. The component data is then recalibrated after the equipment is repaired. This mechanism ensures the robustness of data processing in the complex environment of the port area.

[0033] Through the above implementation process, the variable-gain Kalman filter effectively adapts to the unique traffic patterns of the port logistics area. The low-frequency components accurately reflect the freight demand cycle dominated by ship operation plans, while the high-frequency components sensitively capture sudden disturbances such as customs inspections and mechanical failures. This separation method provides a precise data foundation for subsequent targeted traffic control, enabling governance strategies to distinguish between normal transportation demand and emergency response.

[0034] Example 3: See Figure 4 In the implementation of Example 3, we will take a high-tech development zone in a provincial capital city as an example. This area is home to 127 technology company headquarters and 3 university research institutes. Traffic flow exhibits a tidal characteristic, with concentrated entry into the zone during the morning peak and concentrated exit during the evening peak. Sudden traffic demands are often generated due to events such as academic conferences and product launches.

[0035] Every day at 6:00 AM, the dynamic route allocator starts operating, receiving high-frequency fluctuation components of traffic flow data decomposed by Kalman filtering. Specifically, when decomposing traffic flow data, Kalman filtering extracts high-frequency fluctuation components containing information about sudden disturbances. The data transmission module in the system accurately transmits these high-frequency fluctuation components to the dynamic route allocator according to a predetermined data transmission protocol and path. After receiving the data, the dynamic route allocator can adjust its routing strategy in real time based on these high-frequency fluctuation information reflecting sudden changes in traffic. For example, it can avoid road sections congested due to sudden situations and replan more efficient routes for vehicles to optimize traffic flow distribution and improve the overall operating efficiency of the road network. For example, at 8:15 AM on a certain Tuesday, an additional 482 vehicles / hour increase was detected due to the holding of a cloud computing summit. This data serves as the input to the hybrid governance system's control quantity, triggering the control quantity update equation: based on the baseline fluctuation level of 297 vehicles per hour, the current control demand value is predicted to increase by 62%. The resource allocation equation then initiated a strategy correction, combining real-time road network load data (at that time, the saturation of the main roads had reached 0.92) to generate the current control strategy correction value—increasing the original road expansion ratio of 30% to 45%.

[0036] The calculation of dynamic route allocation coefficients follows specific rules: in: Indicates the dynamic route allocation coefficient. This represents the remaining capacity ratio of the current road segment (e.g., a value of 0.28 was measured in a certain calculation). This represents the propagation speed of the wave component (typically 0.83 km / min in this region). It is the diversion capacity index of adjacent road networks (calculated as 1.7 based on the saturation of parallel roads). The historical success rate for passenger diversion is 0.63 (statistical value for this period). This occurred during Thursday morning rush hour due to a subway malfunction. When it drops to 0.15, The value decreased accordingly from 0.41 to 0.29.

[0037] Based on the allocation coefficient, a control rule is generated: when When the value is greater than 0.4, road capacity adjustment will be activated first; when the value is less than or equal to 0. When <0.4, a hybrid strategy is adopted. When the value is less than 0.2, the focus is on traffic guidance. In one implementation, this was applied to the 613 vehicles that fluctuated in volume detected at the end of an artificial intelligence conference. =0.38), the system allocates 68% of the control volume (417 vehicles) to the road capacity adjustment unit: extending the green light time of the Science and Technology Avenue outbound direction by 42 seconds and temporarily opening 3 emergency lanes; the remaining 32% (196 vehicles) is allocated to the traffic guidance strategy unit: pushing the "detour via Chuangye Road" option through variable message signs and marking 4 preferred routes in the navigation APP.

[0038] The operation of the top-level road network stability model is manifested in the optimization of the volatility index. The model continuously monitors the state dispersion of 36 road network nodes. When the dispersion of the Software Park area exceeds the warning value of 0.75 during Friday evening rush hour (the normal threshold is 0.6), the management strategy is automatically adjusted: two mobile traffic lights originally located in the financial district are relocated to the entrance of the Software Park, causing the dispersion in that area to drop back to 0.58 within 18 minutes. The objective function is minimized through constraints, such as stipulating that the coefficient of variation of traffic flow at any node must not exceed 0.35 for 5 consecutive minutes. The operating mechanism of the top-level road network stability model is to comprehensively consider key parameters such as traffic flow, vehicle speed, and saturation of each road segment in the road network, as well as the changing trends of these parameters over time, to assess the stability of the entire road network. In terms of model architecture, its input layer receives various real-time traffic data collected from the underlying traffic monitoring equipment. The intermediate layer uses specific algorithms to fuse, analyze, and process this data, such as using machine learning algorithms or mathematical models to mine the potential relationships and patterns between data. The output layer outputs indicators and evaluation results reflecting the stability of the road network. When the road network stability index falls below a set threshold, the model can issue a timely warning, prompting traffic management departments to take corresponding measures to maintain the stable operation of the road network, such as implementing traffic diversion and adjusting traffic light timing.

[0039] The underlying resource cost model adopts a four-dimensional cost accounting architecture. The initial investment cost accounting for the governance facilities includes the procurement and installation costs of equipment such as intelligent traffic light sets (single set cost 230,000 yuan), geomagnetic detectors (unit price 18,000 yuan), and variable message signs (120,000 yuan each). When calculating operation and maintenance costs, expenses such as power consumption of the signal system (78 kWh per day) and equipment inspection (3 person-times per week) are included. Facility replacement costs are depreciated on a five-year replacement cycle. Traffic delay penalty costs are calculated based on real-time vehicle speed deviations. When the actual vehicle speed on a main road is 15 km / h lower than the expected value for 20 minutes, the system automatically generates a penalty cost item of 37 yuan per minute.

[0040] The calculation of regional comprehensive benefits adopts a benefit aggregation mechanism. The benefits of traffic management include the converted value of improved traffic efficiency (e.g., if an optimization reduces the average commute time by 9 minutes, the daily benefit is 140,000 yuan based on the average hourly wage in the region). Environmental benefits are converted through exhaust emission reduction: when 182 kilometers of ineffective driving mileage is reduced during the morning rush hour, it corresponds to a reduction of 1.7 kilograms of nitrogen oxide emissions, which is converted into economic value according to environmental standards. These benefits, after deducting the four types of costs, form a comprehensive benefit value, which serves as the optimization target of the underlying model.

[0041] During resource allocation, the system establishes a priority mechanism. For the special fluctuations caused by the start of the university semester, the incentive strategy weight for education industry clusters is increased to 0.6. When the system detects a need for emergency vehicle access in the medical industry park, the road capacity adjustment unit immediately activates the "green channel" mode. In one implementation, to ensure cold chain transportation for biopharmaceutical companies, the system increased the dedicated truck phase duration at 12 intersections around the pharmaceutical park by 22% during the morning rush hour, while simultaneously inducing other vehicles to detour.

[0042] Through the above implementation process, the hybrid governance system achieves dynamic and precise regulation. The road capacity adjustment unit changes the distribution of traffic capacity through physical facilities, while the traffic guidance strategy unit adjusts traffic flow allocation by using information guidance. The two work together to address different types of traffic fluctuations. The hierarchical optimization model ensures the scientific nature of the governance strategy from both system stability and economic dimensions, ensuring that traffic resource allocation meets both real-time needs and the principle of long-term benefits.

[0043] Example 4: Taking the core transshipment area of ​​a national-level logistics hub city as an example. This area includes three major transportation hubs: an international land port, a railway marshalling yard, and an air cargo center, handling more than 200,000 tons of cargo daily. The transportation constraints need to balance the characteristics of multimodal transport with the needs of urban commuting.

[0044] Collaborative constraints are dynamically managed through parameterized configuration. Road capacity adjustment constraints set node capacity thresholds: the minimum capacity threshold for marshalling yard exit roads is maintained at 400 vehicles / hour (ensuring basic collection and distribution functions), and the maximum capacity threshold is set at 1850 vehicles / hour (to avoid road overload). On a Thursday morning at 10:00 AM, when the detected flow of container trucks reached 1820 vehicles, the system automatically triggered a flow restriction mechanism, closing two secondary entrances to control the flow within the threshold.

[0045] Table 1. Configuration of the release cycle of guidance information in logistics hub areas. When the demand for road transport surged on Tuesday afternoon due to railway train delays, the system automatically switched the multimodal transport connection information release cycle from the normal 30 minutes to an event response level of 5 minutes, and pushed 7 temporary connection routes to the freight APP.

[0046] The hybrid governance system has a 90-second coordination time limit for response constraints. In an emergency drill, a tanker truck leak occurred at the air cargo center. The system completed the response chain within 38 seconds: in the first 15 seconds, the road capacity adjustment unit closed three lanes in the affected area and simultaneously activated the backup cargo lane; in the next 25 seconds, the traffic guidance strategy unit sent detour instructions to 500 trucks in the surrounding area; and in the remaining 22 seconds, the system coordinated with the bus dispatching system to adjust three bus routes. The entire response time was 82 seconds, which is below the threshold requirement.

[0047] The load change rate threshold is set at ±15% / hour. When a 17% increase in hourly traffic flow in the cold chain logistics area is detected during Wednesday morning rush hour, the system immediately initiates tiered control: firstly, temporary parking lanes are opened to increase traffic capacity by 12%; if the change rate still reaches 16%, 50% of vehicles are further guided to detour via the dedicated fresh produce lane. The state dispersion threshold is set at 0.65 (out of 1.0). During a cross-border e-commerce promotion, the dispersion in the daily necessities storage area rose to 0.68. The system dynamically allocated 17% of trucks to the night lane, reducing the dispersion to 0.61.

[0048] A real-time traffic supply calculation model was established: the basic supply capacity of the land port area was fixed at 1200 standard train trips / hour, and the adjustment capacity of the mixed management system fluctuated according to the ship berthing schedule. On a certain Friday when receiving a large cargo ship, the system monitored a basic demand of 1850 train trips and an adjustment demand of 430 train trips, requiring a total supply of 2280 train trips. Balance was achieved by implementing three measures: temporarily activating the railway yard access (+280 train trips), opening a tidal flow lane (+110 train trips), and guiding 40 freight cars to stagger their travel times (-40 train trips).

[0049] The system configures traffic resource ratios for each industrial zone as follows: Heavy-duty lanes account for ≥35% in the international trade zone; loading and unloading berths are ≥20 per square kilometer in the e-commerce warehousing zone; and dedicated lane coverage in the cold chain logistics zone is ≥60%. In one implementation, for a newly established biopharmaceutical industrial park, the system automatically matched the characteristics of the pharmaceutical industry: increasing the proportion of temperature-controlled transport lanes to 28%, adding 12 dedicated parking spaces for biological products within a 3-kilometer radius, and allocating separate passage periods for hazardous materials transport.

[0050] When multiple constraints are triggered simultaneously by a typhoon, the system will handle them according to the preset levels: the road network stability constraint (level 1) will forcibly reduce the intensity of port operations by 30%; the system supply and demand balance constraint (level 2) will activate the emergency storage yard to accept backlogged goods; and the economic structure constraint (level 3) will temporarily allow general cargo vehicles to use the dedicated lane for hazardous chemicals.

[0051] The system continuously monitors the implementation of constraints through 78 roadside units. When road construction causes the marshalling yard exit capacity to be below the minimum threshold for 2 consecutive hours, the system automatically generates a three-level alarm: a primary alarm (below the threshold by 5%) prompts the duty officer to check; a secondary alarm (below the threshold by 10%) triggers the navigation system to avoid the road section; and a senior alarm (below the threshold by 15%) directly triggers the reconfiguration of the governance strategy.

[0052] The basic constraints of road capacity and guidance strategies ensure the basic operation of the system, while hybrid response and road network stability constraints enhance system resilience. Supply-demand balance and economic structure constraints achieve a deep alignment between resource allocation and regional development. This multi-dimensional constraint system provides an operational framework for the refined management of complex transportation systems.

[0053] Example 5: See Figure 5 This paper takes the island-loop transportation corridor of a coastal tourist city as an example. This area connects the airport, cruise terminal, resort hotel clusters, and core scenic spots. During peak tourist season, traffic demand can reach 4.8 times that of the off-season, and it exhibits a significant tidal characteristic of "early arrival and late departure" daily. The cross-layer iterative update solution is initiated at 5:00 AM on the first day of the peak tourist season, with the system inputting the traffic control and allocation instruction set generated the previous night into the hierarchical optimization model.

[0054] A traffic flow simulation platform built based on historical data simulated a commuting scenario from the airport expressway to the hotel area between 7:00 and 9:00 AM. The initial configuration included: two tidal flow lanes, 12 signal priority intersections, and 3 reserved bus lanes. Simulation results showed that the fluctuation index of the coastal road section reached 0.73 (threshold 0.6), and the system automatically marked this section as the focus of optimization.

[0055] The initial resource allocation plan was then transferred to the underlying resource cost model, and the cost accounting module initiated four-dimensional calculations: the tidal flow lane conversion requires the investment of movable barriers (costing 80,000 yuan per kilometer), the signal priority system involves equipment upgrades at 36 intersections (costing 45,000 yuan per point), and the bus lane requires supporting electronic monitoring facilities (costing 23,000 yuan per kilometer). The operating cost calculation includes the cost of mechanical adjustments to the tidal flow lane three times a day (180 yuan per adjustment), as well as the increased energy consumption of the signal system (an additional 62 kWh per day).

[0056] The revised resource cost model incorporates real-time demand data. When the system detects that the first cruise ship arrives early and an additional 832 passengers arrive, it recalculates the control demand: the original plan required an increase of 28% in the airport expressway's control capacity. The model decomposes and transmits the control parameters of the preliminary plan, with the road capacity control unit receiving lane configuration parameters and the traffic guidance strategy unit acquiring detour route data.

[0057] The bidirectional iterative solution was initiated during actual morning rush hour operation. In the first iteration, the top-level model detected a 1.2-kilometer queue (discrepancy 0.71) at the scenic area entrance due to improper signal timing, requesting the bottom-level model to increase the guidance intensity. The bottom-level model's evaluation found that increasing the guidance intensity by 40% would require deploying 3 more mobile information vehicles (increasing daily costs by 2400 yuan), but would reduce delay penalty costs by 12% (approximately 1700 yuan / hour). After a benefit-benefit comparison, the system adopted this suggestion and fed it back to the top-level model.

[0058] When the underlying model calculations showed that the utilization rate of the coastal road's tidal lanes was only 65% ​​of the estimated level, a revised plan was proposed to reduce the configuration of guardrails by 20%. The top-level model re-simulated and verified this: after the lane reduction, the volatility index rose to 0.68 but still met the threshold, so the revision was approved. Simultaneously, the saved 160,000 yuan in equipment costs was reinvested in the scenic area's parking guidance system. At 8:17 AM, a rear-end collision occurred on the cross-sea bridge, and the top-level model detected a sharp increase in the dispersion of the associated road network to 0.79. The system immediately activated its emergency mechanism: the underlying model temporarily lifted the economic constraints and approved the use of helicopter cruise guidance (hourly cost 18,000 yuan), while the top-level model instructed the opening of wartime backup roads. After the accident was cleared, the system automatically restored the normal constraints.

[0059] The final solution reflects a multi-dimensional balance. The airport expressway adopts a dynamic tidal flow lane system: during the morning rush hour, the main road into the city is expanded to five lanes, with navigation systems simultaneously displaying a "detour via Yugang Road" prompt. Tiered guidance is implemented around scenic areas: parking availability information is pushed out 3 kilometers from the entrance, and bus connection reminders are triggered 1 kilometer away. A reservation system is implemented in the cruise terminal area, diverting 30% of traffic flow to nighttime hours. In the resource allocation plan, intelligent traffic light coverage reaches 78%, the number of mobile guidance vehicles deployed is optimized from the planned 9 to 7, and the frequency of emergency lane activation is controlled to an average of 1.2 times per day.

[0060] During the implementation phase, the system maintains continuous calibration. Every 15 minutes, the actual traffic conditions are compared with the model predictions. When it is detected that the morning rush hour in the hotel area dissipates 18 minutes earlier than expected on Friday, the system automatically optimizes the plan for the next day: the tidal lane closure time is moved forward from 9:30 to 9:15, saving mechanical adjustment costs while improving the utilization rate of the bus lane.

[0061] Through multiple rounds of closed-loop iterations, the tourism corridor achieved a road network fluctuation index of 0.58 on the first day of the Golden Week holiday (target value 0.6). The solution not only ensured the safe operation of the airport expressway with a maximum capacity of 1,900 vehicles per hour, but also kept investment in congestion mitigation facilities within 85% of the regional transportation budget, thus forming a sustainable congestion management model.

[0062] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0063] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A traffic congestion management optimization algorithm based on regional population size and economic structure, characterized in that, include: Acquire population distribution data, economic and industrial data, and real-time traffic flow raw data of the target area, and perform multi-dimensional clustering processing on the population distribution data, economic and industrial data, and real-time traffic flow raw data to generate a representative dataset of regional traffic characteristics. Based on the representative dataset of regional traffic characteristics, a set of traffic control and allocation instructions is generated. A dynamic traffic state hierarchical optimization model is constructed based on the traffic control and allocation instruction set. The dynamic traffic state hierarchical optimization model includes a top-level road network stability model and a bottom-level resource cost model. The collaborative constraints of the top-level road network stability model and the bottom-level resource cost model are determined. The traffic control and allocation instruction set is input into the top-level road network stability model and the bottom-level resource cost model. Cross-level iterative update is performed based on the collaborative constraints to solve the problem and output the optimal traffic management strategy configuration scheme.

2. The traffic congestion management optimization algorithm based on regional population size and economic structure according to claim 1, wherein generating a traffic control and allocation instruction set based on the regional traffic feature representative dataset includes: The representative dataset of regional traffic characteristics is subjected to dynamic gain Kalman filter decomposition to separate the high-frequency fluctuation component of traffic flow and the low-frequency trend component of traffic demand. A hybrid governance system is selected by road capacity adjustment and traffic guidance strategies. The low-frequency trend component of traffic demand is used as the road network load benchmark value, and the high-frequency fluctuation component of traffic flow is dynamically allocated within the hybrid governance system to generate a traffic control and allocation instruction set.

3. The traffic congestion management optimization algorithm based on regional population size and economic structure according to claim 1, characterized in that, The step of performing multi-dimensional clustering processing on the population distribution data, economic industry data, and real-time traffic flow raw data to generate a representative dataset of regional traffic features includes: The population distribution data, economic industry data, and real-time traffic flow raw data are normalized to generate a standardized set of traffic parameters. The silhouette coefficient method is introduced to calculate the multidimensional clustering quality assessment index; Determine the silhouette coefficients for different numbers of clusters and construct silhouette coefficient variation curves. When the silhouette coefficient variation curves reach a preset stable threshold, output the optimal number of clusters. Based on the optimal number of clusters, initial cluster centroids are selected from the standardized traffic parameter set. The weighted Euclidean distances from all standardized traffic parameters to the initial cluster centroids are calculated. Each standardized traffic parameter is assigned to the cluster corresponding to the minimum weighted Euclidean distance. The cluster centroids are iteratively updated until the centroid offset is lower than the convergence threshold, so as to generate a representative dataset of regional traffic features.

4. The traffic congestion management optimization algorithm based on regional population size and economic structure according to claim 2, characterized in that, The step of performing dynamic gain Kalman filter decomposition on the representative dataset of regional traffic features includes: The regional traffic feature representative dataset is input into a variable gain Kalman filter for state prediction, and a mathematical model of regional traffic state evolution is established. Configure the time update equation and state correction equation of the mathematical model for the evolution of traffic state in the region; Based on historical traffic conditions, the current traffic flow observation is predicted. The current traffic flow observation is then input into the state correction equation for residual calibration, and the calibrated traffic state estimate is output. The calibrated traffic state estimate is input into the time update equation to predict the prior estimate of traffic state for the next period, so as to separate the high-frequency fluctuation component of traffic flow from the low-frequency trend component of traffic demand.

5. The traffic congestion management optimization algorithm based on regional population size and economic structure according to claim 2, characterized in that, The step of dynamically allocating the high-frequency fluctuation components of traffic flow within the hybrid management system to generate a traffic control and allocation instruction set includes: The high-frequency fluctuation component of the traffic flow is used as the control quantity input of the hybrid governance system into the dynamic routing distributor. Establish a mathematical model for traffic control and allocation based on Kalman prediction; Configure the control quantity update equation and resource allocation equation of the traffic control and allocation mathematical model; Based on the traffic fluctuation components of the previous period, the current control demand value is predicted, and the current control demand value is input into the resource allocation equation for strategy correction to generate the current control strategy correction value. The current control strategy correction value is input into the control quantity update equation to predict the traffic control allocation quantity for the next cycle. Calculate the dynamic route allocation coefficients; Based on the dynamic routing allocation coefficient, a traffic control allocation rule is generated. According to the traffic control allocation rule, the predicted traffic control allocation amount for the next cycle is allocated to the road capacity adjustment unit and the traffic guidance strategy unit to generate a traffic control allocation instruction set.

6. The traffic congestion management optimization algorithm based on regional population size and economic structure according to claim 2, characterized in that: The top-level road network stability model takes minimizing the road network fluctuation index as its objective function. The underlying resource cost model includes four sub-models: initial investment cost of governance facilities, operation and maintenance cost, facility renewal cost, and traffic delay penalty cost. The underlying resource cost model takes maximizing the overall regional benefits as its objective function. The overall regional benefits consist of the traffic management revenue minus the initial investment cost, operation and maintenance cost, facility renewal cost, and traffic delay penalty cost of the management facilities, plus the environmental benefits of improved traffic efficiency.

7. The traffic congestion management optimization algorithm based on regional population size and economic structure according to claim 2, characterized in that, The collaborative constraints include: Road capacity adjustment constraints: The traffic capacity of road network nodes is between the preset minimum traffic threshold and the maximum carrying capacity threshold; Traffic guidance strategy constraints: The frequency of guidance information release meets the preset release cycle requirements.

8. The traffic congestion management optimization algorithm based on regional population size and economic structure according to claim 7, characterized in that, The collaborative constraints also include: Hybrid governance system response constraints: The coordinated response time of the road capacity adjustment unit and the traffic guidance strategy unit is lower than the system response threshold; Road network stability constraints: The road network load change rate and the road network state dispersion shall not exceed the preset road network stability boundary.

9. The traffic congestion management optimization algorithm based on regional population size and economic structure according to claim 8, characterized in that, The collaborative constraints also include: System supply and demand balance constraint: The total regional transportation supply equals the sum of the basic transportation demand and the regulation capacity of the mixed governance system; Economic structure constraints: The allocation ratio of transportation resources in each industrial cluster is consistent with the characteristics of the regional industrial economic structure.

10. The traffic congestion management optimization algorithm based on regional population size and economic structure according to claim 1, characterized in that, The step of inputting the traffic control and allocation instruction set into the top-level road network stability model and the bottom-level resource cost model, and performing cross-layer iterative update and solution based on the collaborative constraints, includes: The traffic control and allocation instruction set is input into the top-level road network stability model to perform road network fluctuation analysis and generate a preliminary governance resource allocation plan. The preliminary governance resource allocation scheme is input into the underlying resource cost model for cost-benefit correction, generating a corrected resource cost model. The revised resource cost model integrates real-time traffic demand data to calculate the regulation requirements of the hybrid governance system. The regulation parameters corresponding to the preliminary governance resource allocation scheme are input into the road capacity adjustment unit model and the traffic guidance strategy unit model, respectively. With the goal of maximizing the comprehensive regional benefits, the model is solved in a two-way iterative manner under the cooperative constraints to output the optimal traffic governance strategy allocation scheme.

Images