Bus priority traffic scheduling method, device and equipment based on traffic entropy

Abstract

The invention discloses a bus priority traffic scheduling method, device and equipment based on traffic entropy, and the method comprises the steps: generating arrival time information through obtaining bus position data, and building a time credit line mechanism based on the punctuality rate and the historical application frequency to carry out the differential resource distribution of each line; executing multi-line resource competition analysis to determine an operation load index, and formulating a priority execution strategy based on quota constraint; a traffic entropy theory is introduced, the chaos degree caused by signal adjustment is calculated through speed variance and queuing length change, and an entropy-constrained green light timing scheme is generated; analyzing a road network sweep effect formed by traffic disturbance propagation, and constructing an entropy decline control mechanism through game equilibrium analysis and an entropy propagation model; a road network coordinated regulation and control strategy is formulated, unified management and control of the whole network are realized through a hierarchical coordination architecture and time load liquidation, a global scheduling strategy based on entropy optimization is finally formed, and intelligent evolution control of a traffic system from disorder to order is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent traffic control, in particular to a bus priority traffic scheduling method, device and equipment based on traffic entropy. BACKGROUND

[0002] With the continuous expansion of city size and the continuous growth of motor vehicle ownership, traffic congestion has become an important factor restricting the development of the city. Public transportation is considered as a key way to solve the urban traffic problem due to its high carrying efficiency and low per capita emission characteristics. As an important technical means to improve the level of public transport service, bus priority signal control can reduce the number of stops and delay time of public transport vehicles by providing priority for public transport vehicles at signalized intersections.

[0003] However, in the urban center area where the bus lines are dense, multiple lines often share the same road resources and signalized intersections. When multiple buses arrive at the same intersection at the same time or successively, how to reasonably allocate the limited priority resources becomes a complex technical problem. The existing technology usually adopts a simple strategy of first come first served or fixed priority, ignoring the actual operation status and service demand differences of different lines. In addition, the signal priority measures at a single intersection will have a chain effect on downstream intersections, which may lead to uneven distribution of traffic flow and formation of new congestion points. The traditional method lacks evaluation and control means for such systematic influence, and often appears the phenomenon of "pressing down the gourd to float up the dipper". At the same time, there is a natural contradiction between bus priority and the demand for social vehicle traffic, and how to maintain the stability of the overall traffic order while ensuring bus priority requires more delicate and intelligent control strategies. SUMMARY

[0004] The present application provides a bus priority traffic scheduling method, device and equipment based on traffic entropy, aiming to solve the key technical problems of uneven resource allocation among multiple lines, difficulty in quantifying signal adjustment impact, lack of coordinated optimization among intersections, etc. in traditional bus priority control, through innovative technical means such as time credit management, traffic entropy calculation, game equilibrium analysis, and entropy decreasing control, the present application comprehensively and accurately processes key links such as bus arrival time prediction, time resource dynamic allocation, entropy constrained signal control, and road network disturbance coordinated regulation, and further realizes the dynamic balance optimization of bus priority and traffic order based on entropy theory, and finally establishes an intelligent bus priority control system with entropy value prediction ability and self-adaptive regulation function.

[0005] The first aspect of the present application provides a bus priority traffic scheduling method based on traffic entropy, comprising the following steps: Collecting the trajectory coordinate sequence output by the bus vehicle positioning terminal, and generating arrival time information by performing time prediction processing on the trajectory coordinate sequence; time load evaluation is performed on the arrival time information to generate time credit quotas of each line, a time resource pool is established according to the time credit quotas, and time allocation weights of each line are obtained through dynamic allocation of the time resource pool; Based on the arrival time information, multi-line resource competition analysis is performed to determine line operation load indexes, time allocation requirements are generated according to the time allocation weights and the line operation load indexes, and a time resource allocation scheme is generated through time load balancing processing based on the time allocation requirements; According to the time resource allocation scheme, a priority execution strategy based on quota is formulated, the priority execution strategy is applied to phase adjustment to generate signal changes, traffic entropy value calculation is performed on the signal changes to obtain current traffic confusion degree data, and a green light timing scheme is generated through phase extension analysis on the traffic confusion degree data; Traffic disturbance propagation analysis is performed on the green light timing scheme to generate road network ripple effects, inter-junction game analysis is performed on the road network ripple effects to generate game equilibrium points, confusion degree diffusion parameters are obtained through entropy propagation analysis based on the road network ripple effects, and an entropy decreasing control mechanism is constructed by using the confusion degree diffusion parameters and the game equilibrium points; Full path impact evaluation data are generated in combination with the entropy decreasing control mechanism and the time resource allocation scheme, a chain regulation strategy is formulated according to the full path impact evaluation data, and an inter-junction resource balance scheme is generated according to the chain regulation strategy; A hierarchical coordination mechanism is constructed according to the inter-junction resource balance scheme to generate a comprehensive control scheme, load balancing states are obtained through time load clearing verification on the comprehensive control scheme, a comprehensive optimization strategy is generated based on the load balancing states, and bus priority traffic scheduling is completed.

[0006] The second aspect of the application provides a bus priority traffic scheduling device based on traffic entropy, comprising: A data acquisition module is configured to collect trajectory coordinate sequences output by a bus positioning terminal, and perform time prediction processing on the trajectory coordinate sequences to generate arrival time information; A load evaluation module is configured to perform time load evaluation on the arrival time information to generate time credit quotas of each line, establish a time resource pool according to the time credit quotas, and obtain time allocation weights of each line through dynamic allocation of the time resource pool; A deployment processing module is configured to perform multi-line resource competition analysis based on the arrival time information to determine line operation load indexes, generate time allocation requirements according to the time allocation weights and the line operation load indexes, and generate a time resource allocation scheme through time load balancing processing based on the time allocation requirements; The signal control module is used for formulating a quota-based priority execution strategy according to a time resource allocation scheme, applying the priority execution strategy to phase adjustment to generate signal changes, performing traffic entropy value calculation on the signal changes to obtain current traffic confusion degree data, and performing phase extension analysis on the traffic confusion degree data to generate a green light timing scheme. The game analysis module is used for performing traffic disturbance propagation analysis on the green light timing scheme to generate a road network wave effect, performing intersection game analysis on the road network wave effect to generate a game equilibrium point, performing entropy propagation analysis on the road network wave effect to obtain a confusion degree diffusion parameter, and constructing an entropy reduction control mechanism by using the confusion degree diffusion parameter and the game equilibrium point. The cooperative control module is used for generating full-path impact evaluation data by combining the entropy reduction control mechanism and the time resource allocation scheme, formulating a chain regulation strategy according to the full-path impact evaluation data, and generating an inter-intersection resource balance scheme according to the chain regulation strategy. The comprehensive optimization module is used for constructing a hierarchical coordination mechanism according to the inter-intersection resource balance scheme to generate a comprehensive control scheme, verifying a load balance state by performing time load clearing on the comprehensive control scheme, generating a comprehensive optimization strategy based on the load balance state, and completing bus priority traffic scheduling.

[0007] The third aspect of the present application provides a computer device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the bus priority traffic scheduling method based on traffic entropy disclosed in the first aspect when executing the program.

[0008] The beneficial effects of the present application are embodied in the following points: 1. By establishing a time credit mechanism and a dynamic time resource pool, differentiated credit allocation is carried out according to the punctuality rate and historical priority application frequency of each line, and dynamic optimization configuration of resources is realized through real-time weight adjustment, which changes the traditional fixed allocation mode, so that the lines with poor operating conditions and heavy service loads can obtain more priority protection, while avoiding the situation of excessive occupation of resources by individual lines, and improving the overall operating efficiency of the public transportation system. 2. The concept of traffic entropy is introduced to quantitatively evaluate the degree of traffic chaos caused by signal changes, and the chaos degree data is obtained through the entropy value calculation of speed variance and queue length change, and the interest correlation between intersections is analyzed by game theory. This quantitative evaluation system makes the influence of priority measures can be accurately predicted and controlled, and the managers can predict the possible negative effects before implementing the priority, and control the chaos degree within an acceptable range through the entropy decreasing control mechanism, and realize the balance between priority effect and traffic order. 3. A complete technical system from single-point control to road network coordination is constructed, the propagation path of priority measures is identified through traffic disturbance propagation analysis, road network coordination control strategy and resource balance scheme between intersections are formulated, and unified control of the whole network is realized through hierarchical coordination mechanism, avoiding the problem transfer caused by local optimization, and ensuring the improvement of public transportation service while maintaining the overall operation stability of the road network.

[0009] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF DRAWINGS

[0010] The drawings herein show specific examples of the technical solutions described in the present application, and constitute part of the specification together with the specific embodiments, for explaining the technical solutions, principles and effects of the present application.

[0011] Unless specifically stated or defined otherwise, the same reference signs in different drawings represent the same or similar technical features, and different reference signs may also be used to represent the same or similar technical features.

[0012] Figure 1 is a flowchart of a public transportation priority traffic scheduling method based on traffic entropy according to the present application.

[0013] Figure 2 is a structural block diagram of a public transportation priority traffic scheduling device based on traffic entropy according to the present application.

[0014] Figure 3 is a structural diagram of a computer device according to the present application. DETAILED DESCRIPTION

[0015] In the following description, for purposes of explanation and not limitation, specific details are set forth such as particular architectures, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known methods, devices, circuits, and

[0016] It is to be understood that the terminology "includes", "has", "holds", "contains" and / or similar, as used in the specification and in the following claims, indicates the presence of the described features, integers, steps, operations, elements, and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0017] As used in the specification and in the following claims, the term "if" can be interpreted as meaning "when" or "upon" or "in response to determining" or "in response to detecting" depending on the context. Similarly, the phrase "if it is determined" or "if [the described condition or event] is detected" can be interpreted as meaning "upon determining" or "in response to determining" or "upon detecting [the described condition or event]" or "in response to detecting [the described condition or event]" depending on the context.

[0018] The technical solutions of the embodiments of the present application are introduced as follows.

[0019] As shown in Figure 1 The embodiment of the present application provides a traffic entropy-based bus priority traffic scheduling method, which comprises the following steps S110-S170. In step S110, the trajectory coordinate sequence output by the bus positioning terminal is collected, and the arrival time information is generated by performing time prediction processing on the trajectory coordinate sequence.

[0020] Specifically, the bus positioning terminal outputs a trajectory coordinate sequence in real time, each coordinate point containing four basic elements: longitude, latitude, altitude, and timestamp, and the sampling frequency is set to 1 Hz to ensure the continuity of the trajectory. The positioning terminal integrates a GPS / Beidou dual-mode receiver, an inertial measurement unit, and a vehicle speed pulse interface to generate high-precision trajectory coordinates through a multi-sensor fusion algorithm. The trajectory coordinate sequence is stored in chronological order to form a structured spatiotemporal data stream, and each coordinate point is also attached with auxiliary information such as positioning mode identifier, horizontal precision factor, and number of satellites. When the vehicle is driving in a signal-shielded area, the system automatically switches to an inertial navigation mode to calculate the current position based on the previous trajectory, ensuring the integrity of the coordinate sequence. A real-time cache mechanism for the coordinate sequence is established to retain the last 10 minutes of trajectory data for predictive calculation, and historical data exceeding the time limit are automatically archived. An abnormality detection algorithm for the trajectory coordinates identifies and marks position jumps, speed abnormalities, and other situations, providing data quality identifiers for subsequent processing.

[0021] The collected trajectory coordinate sequence is subjected to time prediction processing. First, the discrete coordinate points are connected by a cubic spline interpolation to construct a continuous driving trajectory curve. The spherical distance and time interval between adjacent coordinate points are calculated to obtain the instantaneous speed sequence of each section. The speed sequence is smoothed by Kalman filtering to eliminate the influence of measurement noise and extract the true speed variation trend. Based on the processed speed data and the remaining distance, an autoregressive moving average model is used to predict the arrival time at each station. The prediction model considers the influence of three factors: road characteristics, historical operation mode, and current traffic state, and improves the prediction accuracy through weighted combination. The generated arrival time information includes three core elements: predicted arrival time, confidence interval upper and lower limits, and prediction credibility score. The information update mechanism ensures that the arrival time is dynamically refreshed immediately after receiving new coordinate points, and finally generates the arrival time information.

[0022] In step S120, the arrival time information is subjected to time load evaluation to generate time credit limits for each line. A time resource pool is established based on the time credit limits, and the time resource pool is dynamically allocated to obtain time allocation weights for each line.

[0023] In some embodiments, the time load evaluation of the arrival time information to generate time credit limits for each line includes: analyzing the arrival time information to obtain arrival time deviation data for each line; generating punctuality rate indicators for each line based on the arrival time deviation data; performing time load evaluation based on the punctuality rate indicators and historical priority application frequencies to obtain load evaluation results; and generating time credit limits for each line based on the load evaluation results.

[0024] The arrival time information is analyzed to obtain the arrival time deviation data of each line. Taking a certain urban trunk line as an example, the standard arrival time is 8:00 am, and the arrival time information shows that the actual expected arrival time is 8:03, so the time deviation is directly calculated as +3 minutes. Through continuous 7-day data collection and statistical analysis, the time deviation data sequence of this line during the morning peak period is [+2, +4, +1, +5, +3, +2, +4] minutes, the average deviation is +3 minutes, and the standard deviation is 1.4 minutes. A classification statistical mechanism of deviation data is established, and the time deviation is accurately classified according to four levels of more than 2 minutes in advance, within ±2 minutes, 2-5 minutes of slight delay, and more than 5 minutes of serious delay. The abnormal value detection algorithm is used to identify and filter extreme deviation data caused by sudden events. When the deviation value exceeds 3 times the standard deviation of the normal range, i.e. ±4.2 minutes, it is marked as an abnormal value and excluded from the statistics. Different types of lines are set with different deviation tolerance standards. The deviation tolerance of the BRT line is set to ±1 minute, the deviation tolerance of the regular bus line is set to ±2 minutes, the deviation tolerance of the community connection line is set to ±3 minutes, and the deviation tolerance of the night shift line is set to ±5 minutes.

[0025] The number of arrivals within the tolerance range in the deviation data is compared with the total number of arrivals to calculate the basic on-time rate index of each line. Based on the deviation data of the urban trunk line [+2, +4, +1, +5, +3, +2, +4] minutes, the number of arrivals within the ±2 minute tolerance range is 2 times (+1 and +2 minutes), and the total number of arrivals within 7 days is 168 times, of which the number of on-time arrivals is 134 times, and the on-time rate is calculated to be 79.8%. A multi-period weighted on-time rate calculation method is established, which divides a day into four periods: morning peak period 7:00-9:00, flat peak period 9:00-17:00, evening peak period 17:00-19:00, and night period 19:00-7:00, with weight coefficients of 1.5, 1.0, 1.5, and 0.8 respectively. The on-time rate of different periods is evaluated by weighted calculation method, and the on-time performance of morning and evening peak periods has greater impact on overall evaluation, highlighting the importance of key periods. An on-time rate level evaluation standard is established, and the line is divided into four levels according to the on-time performance: excellent level above 95%, good level 90%-95%, qualified level 80%-90%, and improvement level below 80%.

[0026] Using the generated punctuality rate index data as the core input of load evaluation, a load evaluation model is established, taking punctuality rate as the measurement standard of line operation quality, and historical priority application frequency as the reflection of line demand for signal priority resources. Taking the urban trunk line with a punctuality rate of 79.8% as an example, combined with the data of historical priority application frequency of 3.5 times per hour, the load evaluation algorithm is used for weighted calculation, the punctuality rate weight coefficient is set to 0.6, and the application frequency weight coefficient is set to 0.4, and the load index of the line is 12.2. The load evaluation algorithm adopts a weighted scoring model: L = 0.4 × Ps + 0.3 × Fs + 0.3 × Cs, wherein Ps is the punctuality rate score equal to the number of punctual arrivals divided by the total number of arrivals multiplied by 100, Fs is the application frequency score equal to the ratio of historical application frequency to standard application frequency multiplied by 100, and Cs is the passenger load rate score equal to the actual passenger load divided by the standard passenger load multiplied by 100. The load level division standard is established, and according to the load index value, the line is divided into four levels of light load 0-5, medium load 5-15, heavy load 15-25, and overload above 25, and the load index of the line is 12.2, which belongs to the medium load range. The load trend analysis algorithm is used to calculate the change of load index in the last 30 days, and the rising, falling or stable trend of load state is identified by comparison and analysis. The load early warning mechanism is established, when the line load index exceeds 20 for 3 consecutive days, the heavy load warning is triggered, and when it exceeds 25 for 7 consecutive days, the overload warning is triggered, and the warning information is automatically sent to the dispatch center. Differentiated load evaluation parameters are set for different line types, the weight coefficient of the fast bus line is set to 0.7, the weight coefficient of the regular line is set to 0.6, and the weight coefficient of the transfer line is set to 0.5, which reflects the operation characteristics of different line types. According to the comprehensive evaluation and trend analysis, the load evaluation result is finally determined.

[0027] The load level and load index from the load assessment results serve as the basis for calculating credit allocation, and a credit allocation algorithm has been established. For example, a line with a load index of 12.2 and a medium load level is used as an example. A base credit of 100 units is used as the starting value. The standard allocation ratio for medium-loaded lines is maintained at 1.0, while an increase factor of 1.2 is set for heavily loaded lines, a decrease factor of 0.8 is set for lightly loaded lines, and an increase factor of 1.5 is set for overloaded lines. A secondary adjustment is made based on the line importance coefficient: 1.3 for trunk lines, 1.0 for secondary lines, and 0.7 for branch lines. The final credit limit for this trunk line is calculated as 100 x 1.0 x 1.3 = 130 units. A credit limit grading system is established, categorizing credit limits into four levels: high credit limit (over 130), standard credit limit (100-130), basic credit limit (70-100), and restricted credit limit (under 70). Different credit limits have different application priority and usage restrictions. A dynamic credit adjustment mechanism was established, updating credit allocation weekly based on the latest load assessment results. Lines with consistently excellent performance received a 10% credit increase, while lines with consistently poor performance received a 10% credit reduction. Credit usage rules were established, stipulating that each priority signal application consumes five credit units. When the credit balance falls below 20 units, non-emergency applications are restricted to ensure the rational use of resources. Through the combined application of the credit allocation algorithm and dynamic adjustment mechanism, precise time credit allocation was achieved for each line.

[0028] All generated time credit data for each route is aggregated into a central resource pool, enabling unified scheduling and allocation of resources across the entire network. For example, for bus routes in a certain city, the credits for each route vary, ranging from 130, 110, 90, and 120, resulting in a total resource pool capacity of 5,500 units. A hierarchical resource pool management mechanism is established, allocating total capacity into three tiers in a 6:3:1 ratio: a core resource tier of 3,300 units for daily priority signal allocation, a buffer resource tier of 1,650 units for peak demand increases, and an emergency resource tier of 550 units for incident and emergency handling. Real-time monitoring parameters are set for the resource pool. When resource utilization exceeds 80% of total capacity, or 4,400 units, a yellow alert is entered; when it exceeds 90%, or 4,950 units, a red alert is entered, automatically initiating resource conservation mode to limit non-emergency requests. A load balancing algorithm is established within the resource pool to monitor resource usage across regions. When resource demand in a particular region surges beyond its allocated capacity, a resource allocation algorithm is used to temporarily redeploy resources from other regions with lower utilization rates. Set up an automatic resource recycling mechanism. When the line operation ends or the resources are not used for 2 consecutive hours, they are automatically recycled to the resource pool for other lines to apply for use.

[0029] The dynamic weight allocation calculation is based on the established time resource pool. A weight calculation model is established, and the current available quota, real-time application frequency and historical use efficiency data of each line are extracted from the resource pool as the three main factors of weight calculation. Take a line as an example, the current available quota of which is 130 units, the real-time application frequency is 2 times every 15 minutes, and the historical use efficiency is 85%. According to the weight proportion of 4:4:2, the comprehensive weight value of the line is 0.142. The comprehensive weight value is calculated by entropy weight method: Wi = (1-Ei) / ∑(1-Ej), wherein Ei is the information entropy of the i-th line, and Ej is the information entropy of the j-th line. A weight normalization processing mechanism is established to calculate and normalize the sum of the weight values of all lines, ensuring that the sum of all weights is equal to 1.0, realizing the complete allocation of resource pool resources. The dynamic weight update frequency is set. Under normal circumstances, the data is re-read from the resource pool and the weight allocation is updated every 15 minutes, and during the traffic peak period, it is adjusted to update every 5 minutes, ensuring the real-time responsiveness of the weight allocation. The weight allocation constraint condition is established to ensure that the minimum weight allocated from the resource pool for each line is not less than 0.01, and the maximum weight is not more than 0.15, preventing excessive concentration of resources to a few lines or some lines from getting no resources at all. When traffic accidents, bad weather and other emergencies occur, additional resources are automatically extracted from the emergency resource layer of the resource pool, and the weight of the affected line is temporarily increased by 20%, and the duration is dynamically adjusted according to the event resolution. Through the coordinated operation of intelligent weight allocation and dynamic update mechanism, the time allocation weight of each line is finally obtained.

[0030] Step S130, based on the arrival time information, the multi-line resource competition analysis is performed to determine the operation load index of each line, the time allocation weight and the operation load index of each line are used to generate the time allocation demand, and the time resource allocation scheme is generated based on the time allocation demand.

[0031] Specifically, a multi-line resource competition analysis model is constructed, and the arrival time information is taken as the input parameter of the competition analysis to quantitatively evaluate the resource competition intensity of different lines at the same space-time node. Taking a major intersection as an example, it is found through analysis that there is a time window overlap of the buses of 3 lines before and after 8:05, the east-west No. 1 line is expected to arrive at 8:04, the south-north No. 2 line is expected to arrive at 8:05, and the east-west No. 3 line is expected to arrive at 8:06, forming a typical resource competition scenario. The resource competition analysis algorithm adopts a multi-dimensional evaluation system, taking the arrival time interval, passenger density, cumulative delay time and line level as the quantitative indicators of competition intensity, and giving them weight coefficients of 30%, 25%, 25% and 20% respectively for comprehensive calculation. The calculation of the operation load index introduces a dynamic parameter model, and the core parameters include real-time passenger load rate, departure interval coefficient, operation period weight and delay accumulation factor. Taking the No. 1 line as an example, the real-time passenger load rate is 85% (85 people / 100 standard capacity), the departure interval is 8 minutes, and the cumulative delay is 3 times a day. Through the weighted algorithm (passenger load rate weight 40%, departure interval weight 30%, delay factor weight 30%), the operation load index is calculated to be 7.2. A load index classification system is established, and the operation load index is divided into four levels: ultra-high load 8.0 and above, high load 5.0-8.0, medium load 3.0-5.0, and low load 3.0 and below. Different levels have different priority weights in resource competition. The resource competition resolution rules clarify the priority sequence of conflict processing, and the ultra-high load line enjoys the highest priority, and the lines with the same level of load are determined in the order of first come first served, delay compensation and passenger priority to allocate resources.

[0032] The time allocation weight is coupled with the operation load index to generate the time allocation demand of each line. For example, the time allocation weight of a certain line is 0.142, the operation load index is 7.2, and the demand calculation model N = W x L x T is adopted, where W is the time allocation weight, L is the operation load index, and T is the standardized time base 10. The time allocation demand of the line is calculated to be 10.2 units. The operation load index calculation formula is L = Σ(wi x fi), where wi is the weight of the ith factor and satisfies Σwi = 1, fi is the standardized function equal to (xi-xmin) / (xmax-xmin), xi is the actual value, xmin is the minimum value, xmax is the maximum value, and the main factors include the passenger load rate (w1 = 0.35), the departure interval deviation (w2 = 0.25), the delay time (w3 = 0.25), and the application frequency (w4 = 0.15). The demand classification management divides the time allocation demand into four levels: the emergency level of 15 units or more, the important level of 10-15 units, the regular level of 5-10 units, and the low priority level of 5 units or less, to ensure the hierarchical management of resource allocation. The conflict identification algorithm scans the demand distribution in the same space-time range, and identifies potential resource conflict points by calculating the time overlap coefficient and the space intersection coefficient. The demand optimization engine performs global optimization under resource constraints, and when the total demand exceeds the available resources, it performs intelligent peak shaving and peak shifting scheduling according to the demand level and line priority. The demand boundary constraint ensures system stability, and the upper limit of the demand of a single line is set to 120% of the maximum service capacity, and the lower limit is not less than 80% of the basic guarantee. The predictive demand analysis is based on historical data and real-time trends to predict demand fluctuations 30 minutes in advance to provide decision support for dynamic resource allocation.

[0033] In some embodiments, the time load balancing processing based on the time allocation demand generates a time resource allocation scheme, including: performing structured processing on the time allocation demand to generate an allocation demand matrix; constructing an inter-line load relationship graph based on the allocation demand matrix; performing balancing processing on the load relationship graph to obtain a time allocation coefficient; and generating a time resource allocation scheme using the time allocation coefficient.

[0034] The time allocation demand is structured to generate the allocation demand matrix. For example, the allocation demand of each line in a certain area during 8:00-9:00 in the morning is 10.2, 8.5, 12.3, 6.8, 15.1, and 9.4 units respectively. A 6x12 demand matrix is constructed in the form of rows representing lines and columns representing time periods, with 12 columns representing a time slice every 5 minutes. The matrix element assignment rule is established, and each matrix element represents the allocation demand of a specific line in a specific time period. The abnormal value beyond the normal range is corrected according to the average value of adjacent time periods. The matrix standardization processing is adopted to normalize the demand value. The maximum-minimum normalization method is used to map all demand values to the interval of 0-1, ensuring the comparability of demand values of different orders of magnitude. The matrix sparsity optimization is established. When the zero elements in the matrix exceed 70%, the compressed storage format is used to reduce memory occupation and improve calculation efficiency. The matrix symmetry test is set up. The symmetry degree of the matrix is calculated to evaluate the balance of demand distribution. When the symmetry degree is less than 0.6, the demand is redistributed.

[0035] The inter-line load relationship graph is constructed based on the allocation demand matrix. Graph theory algorithm is used to take lines as graph nodes and load relationship as connecting edges. The weight value of the edge is determined by the demand difference and time overlap degree of the two lines. The weight calculation formula is W=|D1-D2|×O, where D1 and D2 are the demand values of the two lines, and O is the time overlap degree coefficient. For example, the demand of line 1 in the demand matrix is 10.2 units and the demand of line 2 is 8.5 units. The demand difference of the two lines is 1.7 and the time overlap degree is 80% (0.8). The calculated connecting edge weight is 1.36. The relationship strength classification standard is established, and the relationship between lines is divided into four levels: strong correlation weight above 1.5, medium correlation weight between 0.5 and 1.5, weak correlation weight between 0.1 and 0.5, and no correlation weight below 0.1. Different levels of relationship correspond to different scheduling strategies and coordination mechanisms. Graph clustering algorithm is used to identify line groups with similar load characteristics. The shortest path and connectivity coefficient between nodes are calculated to form the clustering distribution of load relationship. The hierarchical structure of the relationship graph is established, and the graph is divided into three levels: core layer, buffer layer, and peripheral layer. The core layer lines undertake the main transportation task, and the peripheral layer lines provide supplementary services.

[0036] The load relationship graph is balanced to obtain the time allocation coefficient. A balanced objective function is established to minimize the square sum of the line load difference Σ(Li-Lavg)² and maximize the overall efficiency of the system ΣLi / Ti, where Li is the line load, Lavg is the average load, and Ti is the line time consumption. Genetic algorithm is used for iterative optimization, with population size of 50, crossover probability of 0.8, and mutation probability of 0.1. Through multiple generations of evolution, the resource allocation ratio of each line is gradually adjusted. Taking the connection edge with a weight of 1.36 in the graph as an example, after 35 generations of genetic algorithm optimization calculation, the allocation coefficient of line 1 is 0.52, and the allocation coefficient of line 2 is 0.48, and the sum of the two is equal to 1.0 to achieve balanced resource allocation. The balanced constraint condition is established to ensure that the minimum allocation coefficient of each line is not less than 0.1 and the maximum allocation coefficient is not more than 0.6, preventing excessive resource tilt and unfair distribution. The balanced convergence judgment standard is set, and when the improvement amplitude of the fitness function is less than 0.001 for 10 consecutive generations, it is considered to reach the optimal balanced state. The balanced effect evaluation mechanism is established, and the load variance and Gini coefficient before and after the balanced processing are calculated to quantitatively evaluate the improvement degree of the balanced effect.

[0037] The time allocation coefficient is used to generate the time resource allocation scheme. A time period allocation mechanism is established, and 24 hours a day is divided into six periods: morning peak 7:00-9:00, morning flat peak 9:00-12:00, noon peak 12:00-14:00, afternoon flat peak 14:00-17:00, evening peak 17:00-19:00, and night 19:00-7:00. The priority signal time quota of each line is determined according to the allocation coefficient in each period. Taking the morning peak period as an example, the allocation coefficient of line 1 is 0.52, corresponding to 52 seconds of priority green light extension time, the allocation coefficient of line 2 is 0.48, corresponding to 48 seconds, the allocation coefficient of line 3 is 0.35, corresponding to 35 seconds, the allocation coefficient of line 4 is 0.42, corresponding to 42 seconds, the allocation coefficient of line 5 is 0.58, corresponding to 58 seconds, and the allocation coefficient of line 6 is 0.33, corresponding to 33 seconds. A hierarchical execution mechanism of allocation strategy is established, high coefficient lines enjoy priority application right and resource guarantee, medium coefficient lines obtain allocation when resources are sufficient, and low coefficient lines adopt peak shifting allocation strategy to avoid peak conflict. Allocation conflict resolution algorithm is used, when multiple lines apply for priority signal at the same intersection, the service order is determined according to the allocation coefficient from high to low, the coefficient difference of more than 0.1 line has absolute priority, and the coefficient difference of less than 0.1 line adopts time alternation allocation mechanism. The execution rules of the allocation scheme are set, and the allocation time of each line is quantized with 5 seconds as the minimum unit, the duration of single priority signal is not more than 150% of the corresponding time of the allocation coefficient, and not less than 50%.

[0038] Step S140, according to the time resource allocation scheme, formulating the priority execution strategy based on the quota, applying the priority execution strategy to the phase adjustment to generate signal changes, calculating the traffic entropy value of the signal changes to obtain the current traffic chaos degree data, and performing phase extension analysis on the traffic chaos degree data to generate the green light timing scheme.

[0039] Specifically, according to the time resource allocation scheme, the priority execution strategy based on the quota is formulated. The quota-driven algorithm is adopted, and the time-sharing quota of each line is taken as the constraint condition of the execution strategy to perform priority sorting and conflict resolution. Taking a certain intersection in the morning peak period as an example, the No. 1 line is allocated 52 seconds of priority time, the No. 2 line is allocated 48 seconds, and the No. 3 line is allocated 35 seconds. When the three lines simultaneously apply for priority signal, the service order is determined as No. 1→No. 2→No. 3 line from high to low according to the allocated time. A classification system of the priority execution strategy is established, including four basic types of green light extension strategy, red light shortening strategy, phase insertion strategy and phase jump strategy. The green light extension strategy is applicable to the case that the bus is about to arrive and the current phase is green. The extension time is determined according to the allocated quota, and the No. 1 line can be extended for 15-52 seconds, and the No. 2 line can be extended for 12-48 seconds. The red light shortening strategy is applicable to the case that the bus is waiting and the current phase is red. The shortening time is not more than 80% of the allocated quota. The execution constraints of the priority execution strategy are established. The maximum execution time of a single priority strategy is not more than 150% of the allocated quota, the minimum execution time is not less than 5 seconds, and the interval time between two consecutive priority strategies is not less than 30 seconds. A conflict resolution mechanism of the priority execution strategy is set. When the priority demands of multiple lines occur time conflict, the weighted queuing algorithm W_queue=T_allocation×U_urgency is adopted, wherein W_queue is the weighted queuing priority value, T_allocation is the allocated time, and U_urgency is the urgency coefficient.

[0040] The priority execution strategy is applied to the phase adjustment to generate signal changes. With the phase adjustment algorithm, the adjustment amplitude of the phase parameter is calculated according to the priority execution strategy type. Taking the green light extension strategy as an example, when the No. 1 route applies for extending the green light by 25 seconds, the system calculates that the current green light remaining time is 8 seconds, the actual extension time is 25 seconds, and the total time of the adjusted green light is 33 seconds. The timing control of phase adjustment is established to ensure the smoothness and safety of phase switching, and the execution delay of the phase adjustment instruction is controlled within 200 milliseconds. A 3-second yellow light transition time is set during the phase switching process. The range limit of phase adjustment is set, and the single green light extension does not exceed 60 seconds, the red light shortening does not exceed 40 seconds, and the total cycle time change amplitude is controlled within ± 30%. The real-time execution mechanism of phase adjustment is established, and the signal controller executes the phase change immediately after receiving the adjustment instruction, and records the phase parameter changes before and after the adjustment. The phase state monitoring technology is adopted to track the phase change process of the signal light in real time, including the green light time change ΔG, the red light time change ΔR, the cycle time change ΔC and other key parameters.

[0041] In some embodiments, the traffic entropy value calculation of the signal change obtains current traffic chaos degree data, including: sampling the vehicle state sampling data of the vehicle running state caused by the signal change; generating a speed variance parameter and a queue length change parameter based on the vehicle state sampling data; entropy processing the speed variance parameter and the queue length change parameter to obtain an entropy processing result; and generating current traffic chaos degree data based on the entropy processing result.

[0042] A vehicle state sampling system is established, video detectors and geomagnetic sensors are deployed at the entrance and exit positions of the intersection, and the vehicle state sampling data of the vehicle running state within the influence range of the signal change is obtained in real time. A vehicle trajectory tracking algorithm is used to identify and record the position, speed, acceleration and parking state change of each vehicle. Taking the signal change of the east-west green light extension by 13 seconds as an example, the sampling system detects that there are 15 vehicles within a range of 200 meters upstream in this direction, of which 12 vehicles change from the parking state to the driving state, the average starting acceleration is 1.2 m / s², and 3 vehicles maintain uniform speed driving state with an average speed of 35 km / h. A classification recording mechanism for sampling data is established, and the vehicle state is divided into four basic states: acceleration, uniform speed, deceleration and parking, and the number of vehicles, duration and speed change amplitude of each state are recorded. The sampling frequency is set to 10 times per second to ensure that the rapid change process of the vehicle state can be captured. A dynamic adjustment mechanism for the sampling range is established, and the sampling range is determined according to the influence degree of the signal change. The influence range of the green light extension is 300 meters upstream and 100 meters downstream, and the influence range of the red light shortening is 150 meters upstream and 200 meters downstream.

[0043] The speed variance parameter and the queue length change parameter are generated based on the vehicle state sampling data. The statistical analysis method is used to calculate the variance of the state sampling data of 15 vehicles. The speed of the stopped vehicles is recorded as 0, and the speed of the running vehicles is 28-37 km / h. Through variance analysis, the speed variance is 285.6, and the speed standard deviation is 16.9 km / h. The classification standard of speed variance is established, and the speed variance is divided into four levels: low variance 0-100, medium variance 100-300, high variance 300-500, and extremely high variance 500 and above. The speed variance of this intersection belongs to the medium variance level. The queue length detection algorithm is used to count the number of queued vehicles before and after the signal change at each entrance of the intersection. Taking the east-west entrance as an example, there are 12 queued vehicles before the signal change, and after the green light is extended for 13 seconds, the number of queued vehicles is reduced to 3, the queue length change is ΔQ=-9 vehicles, and the queue dissipation rate is 75%. The quantification index of queue length change is established, including the queue growth rate R_growth=ΔQ_positive / Q_initial (where ΔQ_positive is the queue increase, Q_initial is the initial queue length), the queue dissipation rate R_dissipation=ΔQ_negative / Q_initial (where ΔQ_negative is the absolute value of the queue reduction, Q_initial is the initial queue length), the queue fluctuation rate R_fluctuation=|ΔQ| / Q_max (where |ΔQ| is the absolute value of the queue change, Q_max is the maximum queue length during the observation period) and other parameters. The calculation period of the queue length change is set to the signal phase period, and the queue length change parameters are recalculated at the end of each period.

[0044] The entropy value processing result is obtained by entropy value processing on the speed variance parameter and the queue length change parameter. The entropy value processing is based on the application of the second law of thermodynamics in the traffic system, and converts the randomness and uncertainty of vehicle movement into a measurable entropy value index. The calculation of speed entropy adopts the differential entropy method of continuous random variables, and the probability density function of speed distribution is f(v). The speed entropy H_v=-∫f(v)log2f(v)dv, and the integral interval covers the observed speed range. In actual calculation, the continuous distribution is discretized, k speed intervals [v_{i-1},v_i] are divided, the vehicle proportion p_i of each interval is counted, and the discrete entropy formula H_v=-Σ(p_i×log2p_i) is applied, where i is from 1 to k. The greater the speed variance, the more dispersed the probability distribution, and the higher the calculated entropy value, indicating that the difference and unpredictability of vehicle speed are enhanced. The entropy value processing of the queue length change adopts a state transition entropy model, defines a queue state space S={rapid growth, slow growth, stable, slow dissipation, rapid dissipation}, constructs a state transition probability matrix P=[p_{ij}], where p_{ij} represents the probability of transition from state i to state j. The calculation formula of queue entropy is H_q=-ΣΣ(π_i×p_{ij}×log2p_{ij), where π_i is the steady-state probability of state i, which is solved by the stationary distribution of Markov chain. The queue length change parameter determines the state transition probability, the more intense the change, the greater the uncertainty of state transition, and the higher the entropy value. The calculation of the comprehensive entropy value adopts a weighted entropy fusion method, and the total entropy H_total=α×H_v+β×H_q+γ×H_v×H_q is defined, where α and β are linear weight coefficients, and γ is an interaction coefficient, reflecting the coupling effect of speed and queue. The weight coefficients are determined by the entropy weight method, and are adaptively adjusted according to the variation degree of each component entropy. The component with a larger variation degree obtains a higher weight. The normalization processing adopts the maximum entropy normalization method, and H_norm=H_total / H_max, where H_max=log2N is the maximum possible entropy value of the system, and N is the number of system states. The time evolution characteristics of the entropy value are captured by the sliding window method, and the average entropy value and the entropy value change rate in the calculation window are calculated, reflecting the dynamic change trend of traffic chaos degree.

[0045] The current traffic chaos degree data is generated based on the entropy value processing result. The chaos degree conversion formula D=a×H_norm+b is adopted, where D is the chaos degree index, H_norm is the normalized traffic state entropy value, a is the conversion coefficient, and b is the reference adjustment value. The conversion process considers the nonlinear characteristics of the entropy value and the actual meaning of the chaos degree. The traffic entropy value calculation formula is H=α×Hs +β×Hq +γ×Hd, where the speed entropy , pv is the speed interval vehicle ratio, the queuing entropy Hq and the density entropy Hd are calculated in a similar way, the weight coefficients are α = 0.4, β = 0.3, and γ = 0.3, and after normalization, H norm = H / H max. The chaos degree classification standard is established, and the traffic chaos degree is divided into four levels: low chaos 0-80, medium chaos 80-120, high chaos 120-160, and very high chaos 160 and above. Different levels correspond to different traffic management strategies. The multi-dimensional chaos degree evaluation method is adopted, and multiple factors such as speed variance, queue length change, and traffic density are comprehensively evaluated. Each factor is assigned a corresponding weight coefficient according to its impact on traffic chaos degree. The time weight mechanism of chaos degree is established, and the recent chaos degree data has a higher weight, and the historical chaos degree data gradually decays. The dynamic updating mechanism of chaos degree is set, and the chaos degree value is recalculated at the end of each signal cycle to maintain the real-time nature of the chaos degree data.

[0046] The phase length analysis technique is adopted, and the current traffic chaos degree data is used as the basic input for timing optimization. The timing demand evaluation method is adopted, and the timing adjustment intensity is determined according to the chaos degree level. Different adjustment coefficients correspond to different chaos degree levels. The timing demand time of each direction is calculated, and the timing calculation formula T = k × D × w is adopted, where T is the timing adjustment time, k is the adjustment coefficient, D is the chaos index, and w is the direction weight coefficient. The optimization objective of the traffic signal timing is established, and the dual objectives of minimizing the total chaos degree and maximizing the traffic efficiency are used for timing optimization. The timing distribution method is adopted, and the total timing demand is distributed to each phase according to the traffic flow proportion. The green light timing scheme constraint condition is set, and the single green light timing is not less than 45 seconds and not less than 5 seconds. The total cycle timing is not more than 25% of the original cycle. The hierarchical execution strategy of the green light timing scheme is established, and different timing intensities are adopted according to the chaos degree level. The conservative timing is adopted for low chaos degree, and the aggressive timing is adopted for high chaos degree. Taking a major intersection as an example, when the chaos index reaches the medium chaos level, the generated green light timing scheme is that the east-west green light duration is adjusted to 70 seconds, and the north-south green light duration is adjusted to 50 seconds. The timing execution order is determined according to the traffic flow proportion, and the east-west direction is given priority to ensure the timing when the east-west flow is larger, forming a dynamic timing adjustment strategy in different time periods and different directions. Through chaos degree analysis and timing optimization processing, the green light timing scheme is generated.

[0047] In step S150, traffic disturbance propagation analysis is performed on the green light timing scheme to generate road network wave effect, intersection-to-intersection game analysis is performed on the road network wave effect to generate game equilibrium point, entropy propagation analysis is performed based on the road network wave effect to obtain chaos degree diffusion parameter, and entropy reduction control mechanism is constructed by using the chaos degree diffusion parameter and the game equilibrium point.

[0048] Specifically, based on the generated green light timing scheme, traffic disturbance propagation analysis is carried out to identify the propagation mechanism and influence range of signal adjustment in the road network. Traffic disturbance is caused by the continuity of traffic flow. When a road intersection releases more traffic flow through green light delay, these traffic flows will move downstream as a whole, changing the traffic load state of downstream intersections. A tracking mechanism for disturbance propagation is established to determine the propagation path of the disturbance through the topological relationship between intersections and the traffic diversion ratio. The calculation of disturbance intensity considers factors such as initial released traffic volume, propagation distance, and road capacity, forming a decreasing intensity distribution. The propagation time is determined by the distance between intersections and real-time speed. The propagation speed in urban road environment usually fluctuates in the range of 20-40 km / h. The time window characteristics of the disturbance show that it reaches a peak within the initial 2-3 signal periods and then gradually decays to the base level. The propagation characteristics of disturbances on different levels of roads are significantly different. Due to the large traffic volume and continuity, the traffic disturbance on the main road can be transmitted for 3-5 intersections, while the influence range of the branch road is limited to 1-2 intersections due to obvious diversion. Taking the execution of a green light timing scheme as an example, the east-west extension of 25 seconds results in the release of an additional 30 vehicles in that direction. These vehicles reach the downstream intersection 500 meters away after 75 seconds, causing the east-west traffic volume at that intersection to increase by 35%, and the saturation to rise from 0.75 to 0.88. Through systematic propagation tracking and disturbance quantification analysis, the spatial range, intensity distribution, and time evolution of the road network ripple effect are obtained.

[0049] In some embodiments, the inter-intersection game analysis of the road network ripple effect generates a game equilibrium point, including: analyzing the road network ripple effect to obtain the interest correlation relationship of each intersection; constructing an inter-intersection game relationship matrix based on the interest correlation relationship; performing a benefit analysis on the game relationship matrix to obtain a benefit analysis result; and generating a game equilibrium point based on the benefit analysis result.

[0050] By analyzing the influence range, intensity distribution and time evolution characteristics of the wave effect in road network, the interest correlation between intersections is identified. The influence range determines which intersections have direct or indirect interest relations. The influence range of 3-5 intersections on the main road means that these intersections form a close interest community, and the decision-making of each other will produce a chain reaction. The intensity distribution reflects the closeness of interest correlation. According to the 15% decreasing rate, the directly adjacent intersections bear the strongest influence, with a correlation intensity of 0.85, the second-level intersection's correlation intensity drops to 0.72, and the third-level intersection further drops to 0.61. With each additional intersection distance, the interest correlation gradually weakens. The time evolution characteristics reveal the dynamic process of interest influence. The influence begins to appear in the first cycle after the initial adjustment, and reaches the peak in the second and third cycles, when the interest conflict of each intersection is the most intense and needs to be focused on coordination. Then the influence gradually decays. The essence of interest correlation is the redistribution of traffic resources in time and space. Upstream intersections obtain additional travel time resources by extending the green light, and correspondingly transfer the queue and delay to downstream intersections, forming a pattern of this and that. The type identification of correlation relationship is based on the positivity and directionality of influence. When the main traffic flow of upstream and downstream intersections is consistent, it forms a positive synergistic correlation, and both sides can obtain a win-win situation through coordination. When there is a directional conflict in traffic demand, it forms a negative competitive correlation. The improvement of one side must be at the expense of the deterioration of the other side. When there is little exchange of traffic between intersections, it forms a neutral correlation, and the mutual influence can be ignored.

[0051] Based on the obtained interest correlation of each intersection, a game relation matrix is constructed, and the qualitative competitive, synergistic and neutral relationships are converted into quantitative influence coefficients. The matrix is in the form of an N × N square matrix, representing the mutual influence relationship between N intersections. The competitive correlation corresponds to negative elements, indicating negative influence. The synergistic correlation corresponds to positive elements, indicating positive promotion. The neutral correlation corresponds to elements close to zero, indicating weak influence. According to the correlation strength level identified in the interest correlation, the influence coefficient value range is set for intersections of different distances. Directly adjacent intersections have the most direct influence, and their coefficient value range is [-0.85, 0.85]. The influence of the second-level intersection is attenuated after one transmission, and the coefficient range is [-0.72, 0.72]. The influence of the third-level intersection is further weakened, and the coefficient range is [-0.61, 0.61]. The specific value of the matrix element G[i,j] is determined by comprehensive calculation, and the calculation formula is G[i,j]=Sign(Type)×Strength×Distance_Factor, where G[i,j] is the influence coefficient of intersection i on intersection j, Sign(Type) is the relationship type symbol function, -1 for competitive relationship to make the element negative, +1 for synergistic relationship to make the element positive, and 0 for neutral relationship to make the element close to zero; Strength is the basic correlation strength; Distance_Factor is the distance attenuation factor. The diagonal elements G[i,i] are uniformly set to 1, indicating that each intersection has complete control and decision-making power over itself.

[0052] The game relation matrix is analyzed to obtain the benefit analysis result. The total benefit function of intersection i is defined as R[i]=B[i]+Σ(G[j,i]×S[j]), where R[i] is the total benefit of intersection i, B[i] is the basic benefit reflecting the direct traffic improvement brought by the execution of the intersection's own strategy, Σ(G[j,i]×S[j]) is the interaction benefit calculated by adding the influence of all other intersections, G[j,i] is the influence coefficient of intersection j on intersection i, and S[j] is the strategy strength of intersection j. When the element G[j,i] in the game relation matrix is negative, it means that the strategy S[j] of intersection j has a negative impact on intersection i, and the greater the negative value, the more serious the impact. When G[j,i] is positive, it means that it has a positive promoting effect, which helps to improve the traffic conditions of intersection i. The benefit calculation is completed by traversing the ith column of the game relation matrix. The influence coefficients of all intersections on intersection i are extracted, multiplied by the corresponding strategy strength, and then added up to obtain the total influence that the intersection bears under a certain strategy combination. The sparsity of the matrix plays a role in the calculation. Only non-zero elements participate in the operation, and zero elements are directly skipped, improving the calculation efficiency. Different strategy strengths S[j] reflect the intensity of the priority measures taken by intersection j, and the multiplication of the corresponding matrix elements G[j,i] reflects the linear superposition characteristics of the influence.

[0053] Based on the results of the benefit analysis, the equilibrium solution is generated to generate the game equilibrium point. The optimal response function of each intersection is extracted from the benefit analysis results, that is, in the case of given other intersection strategies, the strategy selection that maximizes the benefit R[i] of the intersection is searched from the benefit analysis results. In the initialization stage, each intersection queries the separate action part in the benefit analysis results, selects the strategy that maximizes the basic benefit B[i] of itself as the iteration starting point, and forms an initial strategy combination. In each iteration process, intersection i fixes the current strategy of other intersections unchanged, searches all possible values of S[i] in the benefit analysis results, and finds the strategy S[i]* that maximizes the total benefit R[i] as the new selection. The iteration process continues to advance, and all intersections update the strategy in turn until the strategy adjustment amplitude of all intersections in two consecutive iterations is less than the preset convergence threshold, which is usually set to a change rate of 5%. At this time, it is considered that the candidate equilibrium state is reached. The game equilibrium point solution adopts the Nash equilibrium iteration algorithm: the benefit function R[i]=B[i]-C[i]xS[i]+∑(G[j,i]xS[j]), wherein B[i] is the basic benefit, C[i] is the cost coefficient, S[i] is the strategy variable, and G[j,i] is the influence coefficient of intersection j on intersection i. The candidate equilibrium point is strictly verified, and it is confirmed through the query of the benefit analysis result that under the strategy combination, unilateral adjustment of the strategy of any intersection by ±5 seconds will cause the benefit to decrease, which meets the definition requirement of Nash equilibrium. When the search process finds multiple strategy combinations that meet the equilibrium condition, further comparison is needed, the total benefit R_total=∑R[i] of each equilibrium point is calculated, wherein R_total is the total benefit of the road network, and the strategy combination that maximizes the overall benefit of the road network is selected as the final scheme.

[0054] In some embodiments, the entropy propagation analysis based on the road network ripple effect obtains the chaos degree diffusion parameter, including: performing propagation path tracking on the road network ripple effect to obtain propagation path data; generating a traffic chaos degree diffusion speed based on the propagation path data; obtaining an attenuation law parameter by analyzing the attenuation law of the traffic chaos degree diffusion speed; and generating a chaos degree diffusion parameter according to the traffic chaos degree diffusion speed and the attenuation law parameter.

[0055] The propagation path tracking is used to obtain the propagation path data for the road network contagion effect. The influence range of 3-5 intersections recorded in the road network contagion effect provides the spatial boundary for the path tracking, and the 15% decrease rate of the intensity distribution indicates the main direction of the propagation. Starting from the initial disturbance intersection, the transmission process of the chaos degree is tracked along the actual moving direction of the traffic flow. After the initial chaos degree is generated at intersection A by implementing the priority measures, this chaotic state propagates downstream along with the affected traffic flow. The identification of the propagation path is based on the actual driving trajectory of the vehicle, and the main propagation direction and the secondary branch are determined by analyzing the turning proportion of the traffic flow. The main propagation path extends along the direction of the arterial road, covering the sequence of intersections dominated by the straight traffic flow, and recording the position coordinates and intersection numbers of each node. The secondary path branches at the main intersection and spreads to the adjacent road following the turning traffic flow, marking the branch point position and the split proportion. The time characteristics of the path are obtained by recording the time when the chaos degree reaches each node. The first downstream intersection B receives the propagated chaos degree after 75 seconds, the second intersection C receives it after 152 seconds, and the third intersection D receives it after 234 seconds, forming a complete time sequence. The spatial distance between nodes is determined by the road network topology, with a distance of 500 meters between A and B, 650 meters between B and C, and 550 meters between C and D.

[0056] The chaos diffusion speed is generated based on the distance between nodes and the arrival time in the propagation path data. The diffusion speed is calculated by the formula V = D / T, where V is the diffusion speed, D is the spatial distance between adjacent nodes on the path, and T is the propagation time of the chaos degree between nodes. The diffusion speed of the first segment A-B is calculated as VI = 500 / 75 = 6.67 m / s, which is equivalent to 24 km / h; the diffusion speed of the second segment B-C is V2 = 650 / 77 = 8.44 m / s, which is approximately 30.4 km / h; and the diffusion speed of the third segment C-D is V3 = 550 / 82 = 6.71 m / s, which is approximately 24.1 km / h. The variation in diffusion speed reflects the differences in traffic conditions on different road segments. The higher speed on the B-C segment indicates that this segment is relatively free, and the similar speeds on the A-B and C-D segments indicate similar traffic conditions. The diffusion speed of the branch path is calculated separately, and due to the traffic splitting and turning delay, the diffusion speed of the branch path is usually lower than that of the main path. The time-varying characteristic analysis compares the diffusion speeds at different time periods to form a dynamic speed distribution. The acceleration and deceleration characteristics during the diffusion process are calculated by the speed difference between adjacent road segments, with an initial acceleration of about 0.5 m / s².

[0057] The attenuation law parameters are obtained by analyzing the attenuation law of traffic congestion degree diffusion speed. The differentiated distribution of diffusion speed from 24 km / h to 30.4 km / h reveals the non-uniform characteristics of the diffusion of congestion degree, which requires in-depth analysis of the attenuation law under different diffusion speed conditions. An attenuation analysis framework based on diffusion speed is established, and the road network is divided into high-speed diffusion area, medium-speed diffusion area and low-speed diffusion area, and the attenuation characteristics of each area are studied respectively. The congestion degree in the high-speed diffusion area shows a rapid downward trend, because the traffic flow is more likely to disperse during rapid movement, and the concentration of congestion degree decreases rapidly; the medium-speed diffusion area shows stable exponential attenuation characteristics, which is consistent with the classical diffusion attenuation theory; the attenuation process in the low-speed diffusion area is relatively slow, and the persistence of congestion degree in space is stronger. By comparing and analyzing the spatial distribution data of congestion degree under different diffusion speeds, it is found that there is a positive correlation between diffusion speed and attenuation intensity, and a speed-dependent attenuation model λ(v)=λ0+α·(v-v_avg) is established, where λ0 is the baseline attenuation coefficient, α is the speed sensitivity coefficient, and v_avg is the average diffusion speed. Road environmental factors have a regulating effect on the attenuation process, and the open space of the main road promotes the rapid dissipation of congestion degree, while the narrow environment of the branch road delays the attenuation process, and the difference is reflected by the road width correction coefficient. After systematic attenuation law analysis and model parameter calibration, the attenuation law parameters Λ={λ0,α,γ_w,γ_t} are finally obtained, where λ0 is the baseline attenuation coefficient, α is the speed sensitivity coefficient, γ_w is the road width correction coefficient, and γ_t is the time period adjustment coefficient.

[0058] The complete chaos diffusion parameter Ψ is generated according to the traffic chaos diffusion speed V and the attenuation law parameter Λ, and a parameter synthesis model is adopted: Ψ=G(V,Λ)={V_char,λ_eff,K(d,t),d_c}, wherein V_char is a speed characteristic parameter set, λ_eff is an effective attenuation coefficient, K(d,t) is a diffusion kernel function, and d_c is a critical propagation distance. The chaos diffusion parameter is a comprehensive index system for describing the propagation behavior of traffic chaos in the road network, and the speed characteristics and attenuation characteristics need to be organically integrated. The diffusion speed data V provides the kinetic characteristics of the chaos propagation, including the speed of propagation, acceleration change, speed distribution range, and other time dimension information; the attenuation law parameter Λ describes the spatial evolution law of the chaos intensity, and reflects the quantitative relationship between the chaos and the distance increase. The parameter generation process adopts the methods of feature extraction and parameter synthesis, extracts the steady-state characteristics, transient characteristics, and statistical characteristics from the original speed time series, and forms the speed characteristic parameter set V_char={v_min,v_max,v_mean,σ_v}. The diffusion kernel function K(d,t)=exp(-λ_eff·d)·Θ(v·t-d) comprehensively describes the space-time propagation characteristics of the chaos, wherein the effective attenuation coefficient λ_eff=λ0+α·(v-v_avg) is determined by the attenuation law parameter and the real-time speed. Considering the actual application requirements, the directionality parameter is introduced to distinguish the diffusion differences between the main propagation path and the secondary propagation path, and the main path usually has faster propagation speed and stronger influence range. The dynamic adaptation mechanism enables the diffusion parameter to be adjusted according to the real-time traffic state, and the corresponding parameter configuration is adopted in different traffic scenarios.

[0059] The entropy reduction control mechanism is constructed by using the chaos degree diffusion parameter Ψ and the game equilibrium point. The design goal of the control mechanism is to continuously reduce the overall chaos degree of the road network while maintaining the stable operation of each intersection near the game equilibrium point. The chaos degree diffusion parameter Ψ provides a time window for prediction and intervention. The speed characteristic parameter V_char in Ψ can be used to calculate the time when the chaos degree reaches the downstream intersection, and the effective attenuation coefficient λ_eff can be used to evaluate the influence strength. The game equilibrium point determines the target strategy of each intersection, providing a stable reference benchmark for control. The control strategy adopts a hierarchical architecture. The upper layer formulates the target state of each intersection based on the game equilibrium point, and the lower layer implements predictive control according to the diffusion parameter Ψ. When a high chaos degree is detected at a certain intersection, the entropy reduction control process is immediately started. First, the strategy value of the intersection in the game equilibrium point is queried to ensure that the current control does not deviate from the equilibrium. Then, the diffusion speed in Ψ is used to calculate the time when the chaos degree propagates to each downstream intersection, and pre-adjustment is implemented before the chaos degree arrives. The timing of pre-adjustment is determined by the diffusion kernel function K(d, t) in the diffusion parameter Ψ. If the diffusion speed is 24 km / h, the downstream intersection 500 meters away needs to start preparing 75 seconds in advance. The control strength is dynamically adjusted according to the attenuation law in Ψ. Stronger control measures are needed for close-range intersections, while weaker intervention can be used for long-distance intersections due to natural attenuation. The calculation of the chaos degree diffusion parameter is based on the diffusion equation: diffusion coefficient D = 0.1 × Va², propagation speed v = 0.8 × Vt, attenuation coefficient λ = 1.2 / d², where Va is the average speed, Vt is the actual traffic speed, and d is the propagation distance.

[0060] Step S160, generate full-path impact evaluation data by combining the entropy reduction control mechanism and the time resource allocation scheme, formulate the chain regulation strategy according to the full-path impact evaluation data, and generate the inter-intersection resource balance scheme according to the chain regulation strategy.

[0061] Specifically, the full-path impact assessment data is generated in combination with the entropy-decreasing control mechanism and the time resource allocation scheme. The entropy-decreasing control mechanism provides a dynamic prediction and control strategy for the chaos degree, which can predict the traffic state changes of each intersection at different times. The time resource allocation scheme specifies the priority signal time quota of each line at different time periods, such as 52 seconds for line 1 and 48 seconds for line 2 during the morning peak period. The full-path impact assessment starts from the starting station of the bus line and analyzes the impact at each intersection along the operating route. When line 1 applies for 52 seconds of priority time at intersection A, the entropy-decreasing control mechanism predicts that this measure will increase the chaos degree at intersection B after 75 seconds, and simultaneously calculates the remaining time quota of the line at intersection B. The spatial dimension of the impact assessment covers all intersections passed by the line, and the time dimension extends to the entire operating period. The evaluation indicators include the chaos degree change value of each intersection, the travel time saving of the bus vehicle, the delay increase of the social vehicle, and the fluctuation amplitude of the intersection saturation. The cumulative effect analysis reveals the superimposed impact of multiple priority applications, and when the same line applies for priority at consecutive intersections, the pressure on downstream intersections gradually increases.

[0062] According to the chaos degree change, resource consumption, and cumulative effect of each intersection recorded in the full-path impact assessment data, targeted adjustment strategies are developed. The assessment data shows that the chaos degree of intersection B reaches 156, exceeding the allowable threshold of 140, and the resource utilization rate of intersection C reaches 85%, close to the saturation state, which need to be addressed through adjustment strategies. The core idea of the chain adjustment is to convert single-point local adjustment into multi-point coordinated response, avoiding the transfer and accumulation of problems in the road network. The development of adjustment strategies follows the principle of "source control, process optimization, and end compensation". The strength of the priority measure is controlled at the source intersection where the chaos degree is generated, the impact is slowed down in the propagation process by optimizing the signal timing, and the residual impact is eliminated at the end intersection through compensation measures. The strategy types include time adjustment strategy, spatial diversion strategy, and strength control strategy. The time adjustment strategy postpones the priority application of intersection B by 10 minutes by staggering the use of priority, avoiding the peak value of upstream influence; the spatial diversion strategy opens the auxiliary lane at intersection C to guide 30% of the traffic to use the alternative path; the strength control strategy limits the priority time of intersection A from 25 seconds to 18 seconds, reducing the initial disturbance strength. The linkage mechanism of adjustment ensures the coordinated response of upstream and downstream intersections, forming a linkage adjustment scheme of reducing 7 seconds at intersection A, postponing 10 minutes at intersection B, and diverting 30% at intersection C.

[0063] In some embodiments, the generating an intersection-to-intersection resource balancing scheme according to the chain adjustment strategy comprises: decomposing the chain adjustment strategy by intersection to obtain adjustment requirements of each intersection; performing compensation amount processing based on the adjustment requirements of each intersection to obtain compensation amount data; performing coordinated configuration of the compensation amount data to obtain a coordinated configuration result; and generating an intersection-to-intersection resource balancing scheme according to the coordinated configuration result.

[0064] The overall adjustment requirements in the chain adjustment strategy are decomposed into specific intersections, and the adjustment tasks that need to be performed at each intersection are determined. The chain adjustment strategy determines the overall scheme of reducing 7 seconds of priority time at intersection A, postponing 10 minutes at intersection B, and diverting 30% of the traffic flow at intersection C, which needs to be converted into specific implementation tasks at each intersection. The adjustment requirements of intersection A include: shortening the east-west green light from 25 seconds to 18 seconds during 8:15-8:30, and the reduced 7 seconds need to be compensated by other measures; at the same time, monitor the north-south traffic changes to prevent new congestion points. The adjustment requirements of intersection B are: postpone the original priority application at 8:20 to 8:30, maintain the standard signal timing during the waiting period; prepare for the possible traffic superposition after 8:30. The adjustment requirements of intersection C are: open the east auxiliary lane, set up guide signs to guide 30% of the straight traffic flow to the auxiliary lane; adjust the lane signal lights to ensure the safe passage of diverted vehicles. The quantitative parameters of adjustment requirements include time parameters (start and end time, duration), space parameters (involved lanes, affected directions), and intensity parameters (adjustment amplitude, target flow). The constraint conditions limit the adjustment range of each intersection to ensure that the adjustment does not cause new problems. After the decomposition of the intersection, the adjustment requirements of each intersection are obtained.

[0065] Based on the adjustment requirements of each intersection, the compensation amount is processed to obtain the compensation amount data. The impact of the 7-second reduction in priority time at intersection A on public transportation needs to be quantitatively evaluated, and according to the flow-time relationship calculation, the loss of 7 seconds of green light is equivalent to the passage of 12 vehicles, which needs to be compensated by other measures. The calculation of compensation amount uses the equivalent conversion method, and the effects of different compensation measures are different, 1 second of signal cycle optimization is equivalent to the effect of 0.5 seconds of green light extension, and the opening of auxiliary lane can compensate 15 vehicles / hour of passage capacity. The compensation amount of intersection B postponing 10 minutes is reflected in the time dimension, and additional passage guarantee needs to be provided for the affected public transportation vehicles during the postponement period, and it is calculated that 15 seconds of priority time need to be added as compensation. The compensation amount of intersection C diverting 30% of the traffic flow is determined through capacity analysis, and the auxiliary lane needs to provide at least 300 vehicles / hour of additional passage capacity to accommodate the diverted vehicles. The time distribution of compensation is determined according to the impact characteristics, i.e. immediate compensation is executed synchronously with adjustment, and delayed compensation is implemented after the peak. The compensation intensity is set according to the adjustment degree, i.e. 20% compensation amount for light adjustment, 50% for moderate adjustment, and more than 80% for heavy adjustment.

[0066] The compensation amount data shows that intersection A needs 12 vehicles of passing capacity compensation, intersection B needs 15 seconds of time compensation, and intersection C needs 300 vehicles / hour of capacity compensation, which needs to be arranged in the range of the road network. The coordinated configuration first identifies the source of compensation resources. Intersection D has surplus passing capacity at the same period and can undertake part of the compensation demand of intersection A. The signal cycle of intersection E has optimization space and can provide time support for intersection B. Spatial coordination realizes the transfer of the reduced passing capacity of intersection A to intersection D, which are 800 meters apart, and the traffic can be naturally distributed through guidance. The execution sequence of different compensation measures is arranged by time coordination. The capacity compensation of intersection D starts at 8:15, which is synchronized with the adjustment of intersection A. The cycle optimization of intersection E starts at 8:25, which prepares for the postponement of intersection B. Compensation transfer allows surplus resources to flow between intersections. Intersection F transfers the unused 5 seconds of priority time to intersection B to enhance the compensation effect. The coordination priority is determined according to the correlation strength of the intersection. A-D is a strongly correlated intersection pair that prioritizes coordination, followed by B-E, which is moderately correlated.

[0067] According to the combination and implementation timing of the compensation measures determined in the coordinated configuration result, an executable inter-junction resource balancing scheme is generated. The scheme clearly specifies that intersection A performs intensity control from 8:15 to 8:30, reducing the east-west green time by 7 seconds, while intersection D increases the east-west green time by 5 seconds during the same period to accommodate the transferred traffic. Intersection B postpones the priority application from 8:20 to 8:30, and intersection E saves 10 seconds through signal cycle optimization during this period as a time reserve for intersection B. Intersection C opens the auxiliary lane at 8:15 to cooperate with the guidance measures of intersection F to complete 30% of the traffic distribution task. The execution details of the compensation scheme include specific signal parameter adjustment values, lane opening times, traffic guidance measures, and other operation requirements. The hierarchical structure of the scheme includes core compensation and auxiliary compensation. Core compensation directly corresponds to adjustment demand, and auxiliary compensation provides additional support. The timing arrangement ensures that the compensation effect is synchronized with the adjustment impact to avoid imbalance caused by time difference. The monitoring indicators set the thresholds of traffic, delay, and queue length of each intersection to track the execution effect in real time.

[0068] In step S170, a hierarchical coordination mechanism is constructed based on the inter-junction resource balancing scheme to generate a comprehensive control scheme. The comprehensive control scheme is verified for time load clearing to obtain a load balancing state. Based on the load balancing state, a comprehensive optimization strategy is generated to complete the traffic scheduling of bus priority based on traffic entropy.

[0069] Specifically, the hierarchical coordination mechanism is constructed according to the intersection resource balancing scheme to generate the comprehensive control scheme. The cooperative compensation scheme determines specific measures such as reducing 7 seconds at intersection A to increase 5 seconds at intersection D, delaying 10 minutes at intersection B to optimize the cycle at intersection E, and shunting 30% at intersection C to guide at intersection F. These scattered measures need to form a whole control system through hierarchical coordination. The hierarchical architecture sets three control levels: the regional coordination layer is responsible for the linkage control of adjacent 3-5 intersections to ensure the local coordination of compensation measures; the trunk coordination layer manages all intersections along the main road to maintain the continuity of traffic flow; and the network optimization layer balances the resource allocation of each region and trunk from a global perspective. The regional coordination layer classifies intersections A, D, and G into a control unit, and when intersection A executes the time reduction measure, the layer coordinates intersection D to synchronize the time increase, and monitors the response of intersection G. The trunk coordination layer ensures that intersections A, B, and C on the east-west main road maintain a reasonable signal phase difference, so that even if individual intersections are adjusted, the green wave coordination will not be damaged. When the network optimization layer monitors that the northern region has a congestion trend due to compensation measures, it allocates the surplus resources of the southern region for support. The delivery of control commands follows the hierarchical transmission principle, the network layer formulates the overall goal, the trunk layer decomposes into road segment tasks, and the regional layer transforms into intersection instructions. The uploading of real-time information uses the reverse path, the intersection execution situation is summarized to the regional layer, the regional state is reported to the trunk layer, and the trunk operation is fed back to the network layer.

[0070] In some embodiments, the time load clearing verification of the comprehensive control scheme obtains a load balancing state, including: performing comprehensive influence evaluation on the comprehensive control scheme to obtain a traffic system comprehensive benefit; constructing a load optimization mechanism based on the traffic system comprehensive benefit; performing load clearing processing by using the load optimization mechanism to obtain a clearing processing result; and generating a load balancing state according to the clearing processing result.

[0071] The comprehensive control scheme contains all control measures under hierarchical coordination, and the overall effect of these measures needs to be evaluated rather than the single-point benefit. The evaluation dimensions cover four aspects: public transport service improvement, social vehicle impact, road network operation efficiency, and resource utilization. Public transport service improvement is measured by calculating the punctuality rate improvement, travel time reduction, and passenger waiting time reduction of each line. The punctuality rate of Line 1 increases from 79.8% to 85.2%, and the average travel time shortens by 3.5 minutes. The social vehicle impact assesses the additional delay that private cars, taxis, and other transportation modes bear. The average delay of east-west social vehicles increases by 18 seconds per vehicle, and the average delay of north-south social vehicles increases by 12 seconds per vehicle. Road network operation efficiency is reflected by the saturation change, average speed change, and congestion mileage change of key intersections. The average saturation of major intersections decreases from 0.82 to 0.78, indicating an overall improvement in efficiency. Resource utilization evaluates the consumption and remaining amount of time credit for each line. Line 1 uses 78% of the quota, and Line 2 uses 65%. By weighing the positive and negative impacts, the overall benefit value is calculated using a weighted summation method. The public transport improvement weight is 0.4, the social vehicle weight is 0.3, the road network efficiency weight is 0.2, and the resource efficiency weight is 0.1.

[0072] Based on the comprehensive traffic system benefits evaluated, a dynamic load optimization mechanism is constructed. The comprehensive benefit analysis reveals the load imbalance phenomenon at different intersections and lines, with some intersections having excessive load and others having idle resources. An optimization mechanism is needed to rebalance the load. The objective function of load optimization is set to maximize the overall benefit while minimizing the load difference, and a dual-objective optimization method is used to solve it. The optimization variables include time allocation at each intersection, resource quota for each line, and intensity parameters of compensation measures, which can be adjusted. The constraint conditions ensure that the optimization process does not violate the basic traffic operation rules, such as signal cycle length limits, minimum green light time requirements, and line service level bottom lines. The optimization algorithm uses the gradient descent method, which adjusts the parameters in the direction of the fastest benefit improvement from the current load state. In the iteration process, the marginal benefit of each adjustment is dynamically evaluated, and the optimization stops when the marginal benefit is below the set threshold. The load transfer rule allows high-load nodes to transfer part of the task to low-load nodes, but it needs to meet the requirements of spatial proximity and functional similarity.

[0073] The load balancing state is generated according to the resource usage and load distribution characteristics reflected in the clearing processing result. The clearing result shows that after optimization and allocation, the load index of intersection A decreases from 14.8 to 13.1, and the load index of intersection D increases from 6.9 to 8.2, and the load difference between intersections is significantly reduced. The balance of line resources is achieved through the redistribution of surplus, and line 1 obtains 10 units of allocation from line 2, and line 3 obtains 5 units, and the resource guarantee of each line is more balanced. The quantitative indicators of load balancing include load variance, Gini coefficient, range, etc. The load variance after optimization decreases from 18.5 to 12.3, and the Gini coefficient decreases from 0.28 to 0.21. The quantitative indicators of load balancing state: BI = 1 - σ² / σ²max, where σ² = Σ(Li-Lavg)² / n is the load variance, Li is the load of each intersection, Lavg is the average load, n is the number of intersections, and σ²max is the theoretical maximum variance. BI ∈ [0, 1], the closer to 1, the more balanced. The stability of the balanced state is verified by disturbance test, within the normal fluctuation range, the load distribution remains relatively stable and will not accumulate imbalance. The dynamic balance feature shows that the load state adjusts with the change of traffic demand, and the balance point of the morning peak is different from that of the flat peak period, but both are maintained within an acceptable range. The balance margin is set to reserve adjustment space for sudden situations, and each node maintains a load reserve of 15-20%, avoiding full load operation.

[0074] The load balancing state is generated according to the resource usage and load distribution characteristics reflected in the clearing processing result. The clearing result shows that after optimization and allocation, the load index of intersection A decreases from 14.8 to 13.1, and the load index of intersection D increases from 6.9 to 8.2, and the load difference between intersections is significantly reduced. The balance of line resources is achieved through the redistribution of surplus, and line 1 obtains 10 units of allocation from line 2, and line 3 obtains 5 units, and the resource guarantee of each line is more balanced. The quantitative indicators of load balancing include load variance, Gini coefficient, range, etc. The load variance after optimization decreases from 18.5 to 12.3, and the Gini coefficient decreases from 0.28 to 0.21. The quantitative indicators of load balancing state: BI = 1 - σ² / σ²max, where σ² = Σ(Li-Lavg)² / n is the load variance, Li is the load of each intersection, Lavg is the average load, n is the number of intersections, and σ²max is the theoretical maximum variance. BI ∈ [0, 1], the closer to 1, the more balanced. The stability of the balanced state is verified by disturbance test, within the normal fluctuation range, the load distribution remains relatively stable and will not accumulate imbalance. The dynamic balance feature shows that the load state adjusts with the change of traffic demand, and the balance point of the morning peak is different from that of the flat peak period, but both are maintained within an acceptable range. The balance margin is set to reserve adjustment space for sudden situations, and each node maintains a load reserve of 15-20%, avoiding full load operation.

[0075] Based on the obtained load balancing state, a comprehensive optimization strategy for future operation is formulated. The load balancing state provides an optimal benchmark for current resource allocation, with each intersection load index stabilized in the 10-14 interval, and each line resource usage rate maintained in the reasonable range of 70-85%. The formulation of the optimization strategy aims to maintain and improve this balanced state while improving overall operational efficiency. The strategy framework includes three levels of preventive measures, responsive measures, and improvement measures. Preventive measures avoid imbalance through load prediction and advance deployment. When it is predicted that a certain intersection will have high load in 2 hours, the shunting preparation starts 15 minutes in advance. Responsive measures quickly adjust to real-time load fluctuations. A load warning threshold is set, and when the threshold is exceeded, an emergency deployment program is immediately started. The dynamic strategy library contains optimization schemes for different scenarios, such as weekday morning peak, holiday, and large-scale event. Multi-objective coordination ensures the balanced implementation of public transport priority, traffic efficiency, and resource conservation. The analysis of the load balancing state and the formulation of multi-level optimization measures form a comprehensive optimization strategy, which ultimately completes the public transport priority traffic scheduling.

[0076] To implement the traffic entropy-based public transport priority traffic scheduling method corresponding to the above-mentioned method embodiment, to achieve the corresponding functions and technical effects. Referring to Figure 2 , Figure 2 The structure block diagram of a traffic entropy-based public transport priority traffic scheduling device 200 provided by an embodiment of the present application is shown. For ease of illustration, only the parts related to the present embodiment are shown. The traffic entropy-based public transport priority traffic scheduling device 200 provided by the present embodiment includes: A data acquisition module 201 is configured to collect a trajectory coordinate sequence output by a bus positioning terminal, and perform time prediction processing on the trajectory coordinate sequence to generate arrival time information. A load evaluation module 202 is configured to perform time load evaluation on the arrival time information to generate time credit limits for each line, establish a time resource pool according to the time credit limits, and perform dynamic allocation on the time resource pool to obtain time allocation weights for each line. A deployment processing module 203 is configured to perform multi-line resource competition analysis based on the arrival time information to determine operation load indexes for each line, generate time deployment requirements according to the time allocation weights and the operation load indexes for each line, and perform time load balancing processing based on the time deployment requirements to generate a time resource allocation scheme. A signal control module 204 is configured to formulate a quota-based priority execution strategy according to the time resource allocation scheme, apply the priority execution strategy to phase adjustment to generate signal changes, perform traffic entropy value calculation on the signal changes to obtain current traffic chaos degree data, and perform phase extension analysis on the traffic chaos degree data to generate a green light timing scheme. The game analysis module 205 is configured to perform traffic disturbance propagation analysis on the green light timing scheme to generate a road network wave effect, perform intersection-to-intersection game analysis on the road network wave effect to generate a game equilibrium point, perform entropy propagation analysis based on the road network wave effect to obtain a chaos degree diffusion parameter, and construct an entropy-decreasing control mechanism by using the chaos degree diffusion parameter and the game equilibrium point. The cooperative control module 206 is configured to generate full-path impact evaluation data by combining the entropy-decreasing control mechanism and the time resource allocation scheme, formulate a chain regulation strategy according to the full-path impact evaluation data, and generate an intersection-to-intersection resource balancing scheme according to the chain regulation strategy. The comprehensive optimization module 207 is configured to construct a hierarchical coordination mechanism according to the intersection-to-intersection resource balancing scheme to generate a comprehensive control scheme, perform time load clearing verification on the comprehensive control scheme to obtain a load balancing state, generate a comprehensive optimization strategy based on the load balancing state, and complete bus priority traffic scheduling.

[0077] The traffic entropy-based bus priority traffic scheduling device 200 described above can implement the traffic entropy-based bus priority traffic scheduling method described above. The optional items in the method embodiments described above are also applicable to this embodiment, and will not be described in detail here. The remaining content of the present embodiment can refer to the content of the method embodiments described above, and will not be described in detail in this embodiment.

[0078] As shown in Figure 3 The third embodiment of the present application also provides a computer device, which comprises a memory 301, a processor 302, and a computer program stored in the memory 301 and executable on the processor 302, characterized in that the processor 302 implements the steps of the traffic entropy-based bus priority traffic scheduling method according to the first embodiment of the present application when executing the program.

[0079] The above embodiments are not exhaustive enumeration based on the present application, and there can be a plurality of other embodiments not listed. Any replacement and improvement made without violating the concept of the present application is within the protection scope of the present application.

Claims

1. A bus priority traffic scheduling method based on traffic entropy, characterized in that: include: Collecting the trajectory coordinate sequence output by the bus positioning terminal, and performing time prediction processing on the trajectory coordinate sequence to generate arrival time information; Performing a time load evaluation on the arrival time information to generate a time credit for each line, establishing a time resource pool based on the time credit, and dynamically allocating the time resource pool to obtain a time allocation weight for each line; performing a multi-line resource competition analysis based on the arrival time information to determine an operating load index for each line, generating a time allocation requirement based on the time allocation weight and the operating load index for each line, and performing a time load balancing process based on the time allocation requirement to generate a time resource allocation plan; Formulate a quota-based priority execution strategy based on the time resource allocation plan, apply the priority execution strategy to phase adjustment to generate signal changes, calculate traffic entropy values ​​for the signal changes to obtain current traffic chaos data, and perform phase extension analysis on the traffic chaos data to generate a green light timing plan; Performing a traffic disturbance propagation analysis on the green light timing plan to generate a road network ripple effect, performing an inter-intersection game analysis on the road network ripple effect to generate a game equilibrium point, performing an entropy propagation analysis based on the road network ripple effect to obtain a chaos diffusion parameter, and constructing an entropy reduction control mechanism using the chaos diffusion parameter and the game equilibrium point; generating full-path impact assessment data by combining the entropy reduction control mechanism and the time resource allocation scheme, formulating a chain adjustment strategy based on the full-path impact assessment data, and generating an inter-intersection resource balancing scheme based on the chain adjustment strategy; A hierarchical coordination mechanism is constructed based on the resource balancing plan between intersections to generate a comprehensive control plan, the comprehensive control plan is verified by time load settlement to obtain a load balancing state, and a comprehensive optimization strategy is generated based on the load balancing state to complete bus priority traffic scheduling.

2. The method according to claim 1, characterized in that The performing time load evaluation on the arrival time information to generate time credits for each line includes: Analyze the arrival time information to obtain arrival time deviation data for each line; Generating an on-time rate index for each line based on the arrival time deviation data; Performing a time load assessment based on the punctuality index and the historical priority application frequency to obtain a load assessment result; A time credit for each line is generated based on the load evaluation result.

3. The method according to claim 1, characterized in that The performing of time load balancing processing based on the time allocation requirement to generate a time resource allocation scheme includes: Structuring the time allocation requirements to generate an allocation requirement matrix; Constructing a load relationship map between lines based on the deployment demand matrix; Performing balancing processing on the load relationship map to obtain a time distribution coefficient; A time resource allocation scheme is generated using the time allocation coefficients.

4. The method according to claim 1, wherein The step of calculating the traffic entropy value of the signal change to obtain current traffic chaos data includes: Sampling the vehicle running state caused by the signal change to obtain vehicle state sampling data; generating a speed variance parameter and a queue length variation parameter based on the vehicle state sampling data; Performing entropy processing on the speed variance parameter and the queue length change parameter to obtain an entropy processing result; Current traffic chaos data is generated based on the entropy value processing result.

5. The method according to claim 1, wherein The performing of an inter-intersection game analysis on the road network ripple effect to generate a game equilibrium point includes: Analyze the road network ripple effect to obtain the interest correlation relationship of each intersection; Constructing a game relationship matrix between intersections based on the interest association relationship; Performing a benefit analysis on the game relationship matrix to obtain a benefit analysis result; An equilibrium solution is performed based on the profit analysis result to generate a game equilibrium point.

6. The method according to claim 1, characterized in that The entropy propagation analysis based on the road network ripple effect is performed to obtain the chaos diffusion parameter, including: Tracing the propagation path of the road network and its impact to obtain propagation path data; generating a traffic chaos diffusion speed based on the propagation path data; Performing an attenuation law analysis on the diffusion speed of the traffic chaos to obtain attenuation law parameters; A chaos diffusion parameter is generated according to the traffic chaos diffusion speed and the attenuation law parameter.

7. The method according to claim 1, characterized in that Generating an inter-intersection resource balancing plan according to the chain adjustment strategy includes: Decomposing the chain regulation strategy at intersections to obtain regulation requirements at each intersection; Perform compensation processing based on the regulation requirements of each intersection to obtain compensation data; Performing coordinated configuration between intersections on the compensation amount data to obtain a coordinated configuration result; An inter-intersection resource balancing solution is generated according to the collaborative configuration result.

8. The method according to claim 1, characterized in that The performing time load clearing verification on the comprehensive control scheme to obtain a load balancing state includes: Conduct a comprehensive impact assessment of the comprehensive control plan to obtain the comprehensive benefits of the transportation system; Constructing a load optimization mechanism based on the comprehensive benefits of the transportation system; Performing load clearing processing using the load optimization mechanism to obtain a clearing processing result; A load balancing state is generated according to the clearing processing result.

9. A bus priority traffic dispatching device based on traffic entropy, characterized in that: include: A data acquisition module is used to collect the trajectory coordinate sequence output by the bus positioning terminal, and perform time prediction processing on the trajectory coordinate sequence to generate arrival time information; a load evaluation module, configured to perform a time load evaluation on the arrival time information to generate a time credit for each line, establish a time resource pool based on the time credit, and dynamically allocate the time resource pool to obtain a time allocation weight for each line; an allocation processing module, configured to perform a multi-line resource competition analysis based on the arrival time information to determine an operating load index for each line, generate a time allocation demand based on the time allocation weight and the operating load index for each line, and perform time load balancing based on the time allocation demand to generate a time resource allocation plan; A signal control module is configured to formulate a quota-based priority execution strategy based on the time resource allocation plan, apply the priority execution strategy to phase adjustment to generate signal changes, calculate traffic entropy values ​​for the signal changes to obtain current traffic chaos data, and perform phase extension analysis on the traffic chaos data to generate a green light timing plan; a game analysis module configured to perform a traffic disturbance propagation analysis on the green light timing scheme to generate a road network ripple effect, perform an inter-intersection game analysis on the road network ripple effect to generate a game equilibrium point, perform an entropy propagation analysis based on the road network ripple effect to obtain a chaos diffusion parameter, and construct an entropy reduction control mechanism using the chaos diffusion parameter and the game equilibrium point; a collaborative control module, configured to generate full-path impact assessment data in combination with the entropy reduction control mechanism and the time resource allocation scheme, formulate a chain adjustment strategy based on the full-path impact assessment data, and generate an inter-intersection resource balancing scheme based on the chain adjustment strategy; The comprehensive optimization module is used to construct a hierarchical coordination mechanism based on the resource balancing plan between the intersections to generate a comprehensive control plan, perform time load settlement verification on the comprehensive control plan to obtain a load balancing state, generate a comprehensive optimization strategy based on the load balancing state, and complete bus priority traffic scheduling.

10. A computer device, characterized in that: The method comprises a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program and implement the method according to any one of claims 1 to 8 when executing the computer program.

Images