Big data-based intelligent scheduling system and method for vehicle passing through barrier gate

Abstract

The invention relates to the technical field of vehicle scheduling, and discloses a vehicle barrier gate passing intelligent scheduling system and method based on big data. The method comprises the following steps: firstly, dividing an initial vehicle barrier gate traffic data set to obtain a common vehicle barrier gate traffic data set and a non-common vehicle barrier gate traffic data set, and then preprocessing; secondly, constructing a neural network, and outputting a predicted value according to the common vehicle barrier gate passing data set; establishing a vehicle barrier gate passage target function according to the predicted value, solving the vehicle barrier gate passage target function by using a multi-strategy fused Gibshihi shark optimization algorithm to obtain a global optimal solution, and generating a vehicle barrier gate passage scheduling scheme; and finally, adjusting a vehicle barrier gate passing scheduling scheme in real time by using the vehicle barrier gate passing data set which is not commonly used. By analyzing and processing the vehicle barrier gate passing data, the purpose of intelligent scheduling of vehicle barrier gate passing is achieved, and the method is accurate and objective.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle scheduling, and specifically to an intelligent scheduling system and method for vehicle gate access based on big data. Background Art

[0002] Chinese Patent CN115641750B discloses a Beidou-based ship navigation scheduling method and system. The method specifically includes using a Beidou communication device to obtain the number of ships and ship position information in a waterway, dividing the waterway into several regions using the Beidou communication device, calculating the congestion density in several regions according to the number of ships and ship position information to obtain the congestion density in the waterway; then obtaining the ship flow and average ship speed in the waterway, combining the congestion density, establishing a waterway state model, and outputting a waterway state result; setting a state threshold, comparing the waterway state result with the state threshold to determine the navigation state, and completing ship navigation scheduling. This invention only conducts scheduling through congestion density, and the effectiveness and applicability of the scheduling method need to be improved.

[0003] Traditional vehicle gate access scheduling methods often lead to vehicle congestion and even vehicle accidents because vehicles need to be detected; at the same time, due to the lack of technologies such as big data and artificial intelligence, dynamic real-time scheduling cannot be achieved during the vehicle gate access scheduling process, and there is room for improvement in shortening vehicle congestion time and improving traffic capacity. Summary of the Invention

[0004] In view of the problems in the related art, the present invention provides an intelligent scheduling system and method for vehicle gate access based on big data to overcome the above-mentioned technical problems existing in the existing related technologies.

[0005] To solve the above technical problems, the present invention is implemented through the following technical solutions: The present invention is an intelligent scheduling method for vehicle gate access based on big data, including the following steps: S1. Obtain the access records of vehicles entering and leaving the gate to obtain an initial vehicle gate access data set, identify and preprocess the common vehicle gate access data set and the uncommon vehicle gate access data set in the initial vehicle gate access data set to obtain a processed vehicle gate access data set; S2. Based on the processed vehicle gate access data set, predict the vehicle gate access flow and vehicle gate anomalies respectively based on neural networks and time series to obtain a vehicle gate access flow prediction value set and a vehicle gate anomaly prediction value set; S3. Based on the predicted vehicle barrier passage flow value set and the predicted vehicle barrier anomaly set, establish a vehicle barrier passage objective function based on the principle of the best barrier passage capacity, and use the multi-strategy fusion Genghis Khan shark optimization algorithm to solve the vehicle barrier passage objective function to obtain the global optimal solution; S4. Generate a vehicle barrier passage scheduling plan according to the global optimal solution, and introduce the set of infrequently used vehicle barrier passage data to adjust the vehicle barrier passage scheduling plan in real time to obtain the final vehicle barrier passage scheduling plan.

[0006] The invention obtains the initial vehicle barrier passage data set, identifies the frequently used vehicle barrier passage data and the infrequently used vehicle barrier passage data, and preprocesses the frequently used vehicle barrier passage data; this method only processes the frequently used data through data partitioning, reducing the data processing time-consuming. After data preprocessing, the data validity and accuracy are improved, and the subsequent prediction error is reduced; secondly, the vehicle barrier passage flow and vehicle barrier anomalies are predicted based on neural networks and time series respectively to obtain the prediction results. Since there are few vehicle barrier anomalies and they do not have good predictability, it is easy to find the change law according to the time series and the cumulative flow, improving the prediction effect; the LSTM recurrent neural network overcomes the problems of gradient disappearance and gradient explosion, can determine the time lag length, and has good prediction performance; then establish a vehicle barrier passage objective function, and use the multi-strategy fusion Genghis Khan shark optimization algorithm to solve the vehicle barrier passage objective function to obtain the global optimal solution; the objective function models the optimization problem, which is convenient for algorithm optimization. The multi-strategy fusion Genghis Khan shark optimization algorithm simulates the predation and survival behaviors to find the optimal solution. Compared with the traditional algorithm, it avoids falling into the local optimal solution during the iteration process, speeds up the convergence speed, improves the population quality and exploration ability during the iteration process, shortens the vehicle congestion time, and effectively improves the vehicle barrier passage efficiency; finally, generate a vehicle barrier passage scheduling plan and adjust the vehicle barrier passage scheduling plan in real time, considering special vehicles and emergencies, reducing the occurrence of vehicle accidents and improving the reliability of the scheduling plan.

[0007] Preferably, the S1 includes the following steps: S11. Obtain the passage records of vehicles entering and leaving the barrier. The historical passage records of vehicles entering and leaving the barrier include vehicle license plate information, vehicle entry and exit times, vehicle type information, etc., which are recorded as vehicle barrier passage-related information to obtain vehicle barrier passage data; taking the vehicle license plate information in the vehicle barrier passage data as the main body, construct an initial vehicle barrier passage data set , where represents the m th vehicle barrier passage data subset, and the vehicle barrier passage data subset is the vehicle license plate information, and each vehicle license plate information includes other vehicle barrier passage data; S12. Obtain the vehicle entry and exit times in the initial vehicle gate passing data set. Within the time period t, count the number of vehicle entry and exit times. Based on the time period t and the number of vehicle entry and exit times, calculate the vehicle gate passing frequency. Set a frequency threshold. When the vehicle gate passing frequency is greater than the frequency threshold, record the corresponding vehicle license plate information as common vehicles to obtain the common vehicle gate passing data set. Otherwise, record the corresponding vehicle license plate information as uncommon vehicles to obtain the uncommon vehicle gate passing data set. S13. Preprocess the common vehicle gate passing data set, delete the abnormal vehicle gate passing data in the common vehicle gate passing data set, and perform filling. Convert the common vehicle gate passing data set into a processed vehicle gate passing data set. The specific steps are as follows: S131. Within the time period t, divide the time period t by month to obtain the vehicle gate passing month set , where represents the n th vehicle gate passing month; Calculate the average value and mean square deviation of the vehicle gate passing data in each vehicle gate passing month according to the vehicle gate passing month set of the initial vehicle gate passing data set, and delete the vehicle gate passing data corresponding to the maximum mean square deviation and the minimum mean square deviation. Set an average threshold. When the difference between the average value of the vehicle gate passing data and the vehicle gate passing data is less than or equal to the average threshold, the corresponding vehicle gate passing data is recorded as normal vehicle gate passing data. Otherwise, obtain abnormal vehicle gate passing data and delete the abnormal vehicle gate passing data to obtain the processed vehicle gate passing data set. S132. Select any vehicle gate passing month in the vehicle gate passing month set and record it as the vehicle gate passing month , and then select the adjacent vehicle gate passing months of the vehicle gate passing month ; In the processed vehicle gate passing data set, count the missing vehicle gate passing data in the vehicle gate passing month , calculate the average value of the vehicle gate passing data in the adjacent vehicle gate passing months, and use the average value of the vehicle gate passing data to fill the missing vehicle gate passing data in the vehicle gate passing month . Fill the missing vehicle gate passing data in other vehicle gate passing months in turn to obtain the processed vehicle gate passing data set.

[0008] The invention identifies the common vehicle gate passing data and the uncommon vehicle gate passing data, only processes the common data through data division, reduces the data processing time, preprocesses the common vehicle gate passing data, improves the data validity and accuracy, and reduces the subsequent prediction error.

[0009] Preferably, S2 includes the following steps: S21. Obtain the vehicle gate operation information from the processed vehicle gate passing data set. The vehicle gate operation information includes normal operation and faults of the vehicle gate. Divide the processed vehicle gate passing data set into a vehicle gate passing normal operation data set and a vehicle gate passing fault data set according to the vehicle gate operation information. S22. For the vehicle gate passing normal operation data set, train an LSTM recurrent neural network prediction model to predict the vehicle gate passing flow and obtain a vehicle gate passing flow prediction value set. The specific steps are as follows: S221. Obtain the vehicle entry and exit times and the number of vehicle entries and exits in the vehicle gate passing normal operation data set. Use the vehicle entry and exit times as a time series to obtain a vehicle gate passing time series , where represents the th vehicle gate passing time point; calculate the vehicle gate passing flow between each vehicle gate passing time point in turn according to the number of vehicle entries and exits to obtain a vehicle gate passing flow set. Each vehicle gate passing flow in the vehicle gate passing flow set corresponds to a vehicle gate passing time point. S222. Obtain the historical passing records of vehicles entering and exiting the gate again to obtain a vehicle gate passing historical data set. Preprocess the vehicle gate passing historical data set and calculate the vehicle gate passing historical flow to obtain a vehicle gate passing historical time series and a vehicle gate passing flow sample set. Divide the vehicle gate passing flow sample set into a sample training set and a sample test set; set the LSTM recurrent neural network to learn according to the stochastic gradient descent method, set a time sliding window, place the time sliding window on the vehicle gate passing historical time series, and obtain the sample training set included in the vehicle gate passing historical time series within the time sliding window. Input it into the LSTM recurrent neural network in turn and iterate until the LSTM recurrent neural network converges to obtain a trained LSTM recurrent neural network. S223. Place the time sliding window on the vehicle gate passing historical time series to obtain the sample test set included in the vehicle gate passing historical time series within the time sliding window. Input it into the trained LSTM recurrent neural network in turn. Set an accuracy threshold. When the accuracy of the output prediction value is greater than the accuracy threshold, obtain the LSTM recurrent neural network prediction model; otherwise, adjust the weights until the accuracy of the output prediction value is greater than the accuracy threshold. S224, according to the vehicle gate passage time series, input the vehicle gate passage flow set into the LSTM recurrent neural network prediction model, and output the vehicle gate passage flow prediction values ​​in sequence to form a vehicle gate passage flow prediction value set; S23, for the vehicle barrier passage failure data set, set the vehicle barrier failure time, take the vehicle barrier failure time as the time series, and obtain the vehicle barrier passage failure time series ,in Indicates vehicle barrier failure time; calculate the vehicle barrier flow rate between each vehicle barrier failure time, and accumulate it according to the vehicle barrier failure time series to obtain the vehicle barrier cumulative flow rate; use the vehicle barrier failure time series as the horizontal coordinate and the vehicle barrier cumulative flow rate as the vertical coordinate to generate a cumulative flow graph, select the curve segment corresponding to the minimum curve slope in the cumulative flow graph, record it as a sample curve segment, determine the smoothing coefficient, use the least squares method to fit the sample curve segment, obtain the vehicle barrier failure time prediction value, record the vehicle barrier failure time prediction value as the vehicle barrier abnormality prediction value, and form a vehicle barrier abnormality prediction value set.

[0010] The invention predicts the vehicle gate traffic flow and vehicle gate anomalies based on neural networks and time series respectively. Since there are fewer vehicle gate anomalies and they are not very predictive, it is easy to find the changing rules based on time series and accumulated flow to improve the prediction effect. The LSTM recurrent neural network overcomes the gradient vanishing and gradient exploding problems, can determine the time lag length, and has good prediction performance.

[0011] Preferably, S3 comprises the following steps: S31, set the number of vehicles passing through the gate to , combining the vehicle gate flow prediction value set and the vehicle gate abnormality prediction set, based on the principle of optimal gate capacity, determine the average vehicle speed, and obtain the vehicle gate average delay time function and the vehicle gate queue length function; select the minimum value of the vehicle gate average delay time function and the minimum value of the vehicle gate queue length function as the vehicle gate passage objective function, and introduce constraints to meet the vehicle gate opening time is less than or equal to the vehicle gate maximum opening time; S32, introducing the Genghis Khan shark optimization algorithm, and integrating the inferior individual classification strategy and the quasi-adversarial learning strategy to improve the Genghis Khan shark optimization algorithm, to obtain a multi-strategy fusion Genghis Khan shark optimization algorithm, using the multi-strategy fusion Genghis Khan shark optimization algorithm to solve the vehicle gate passage objective function, to obtain the global optimal solution, the specific steps are as follows: S321, the process of solving the objective function of vehicle barrier passage is regarded as a search space, in which there is a Genghis Khan shark population, and the Genghis Khan shark individuals in the Genghis Khan shark population represent candidate solutions of the objective function of vehicle barrier passage; the number of Genghis Khan shark population is set to i , the Genghis Khan shark individual dimension is j , set the current number of iterations to b , initialize the Genghis Khan shark population, and get b +1 iteration i Genghis Khan shark individual location ; During the migration phase, the Genghis Khan shark population updates its position based on the concentration of prey odor, and the odor intensity coefficient is set as , c represents a non-negative constant, represents a random number between the interval [0, 1], then the olfactory intensity , set the b The best Genghis Khan shark individual position at the iteration is , No. b The iteration i Genghis Khan shark individual location , get the b +1 iteration i New location of Genghis Khan shark , at this time, the position To update, ; Get the fitness function value corresponding to the position of the Genghis Khan shark at this time, and record the current best fitness function value and the current worst fitness function value as and , No. i The fitness function value corresponding to the Genghis Khan shark individual is recorded as , introduce the inferior individual classification strategy, calculate pheromone , when the pheromone is less than or equal to 0.3, select Genghis Khan shark individual location and Genghis Khan shark individual location , for position Update, set Represents a random number of 0 or 1. ; S322, Genghis Khan shark population is in the foraging stage, and the step length parameter is set to , based on the best Genghis Khan shark individual position About Location Update again, this time introduce the quasi-adversarial learning strategy, set the upper and lower bounds of the search space to be and , dual value , dual position , calculate the fitness function value corresponding to the dual position, compare it with the current best fitness function, and use the dual position to replace the position , to achieve the position Updates; In the self-protection stage of the Genghis Khan shark population, the Genghis Khan shark individuals use self-protection strategies to position Update to get b +1 iteration i The final position of the Genghis Khan shark individual is obtained, the Genghis Khan shark population is screened, and the next iteration is entered; the maximum number of iterations is set, and when the current number of iterations reaches the maximum number of iterations, the iteration is stopped to obtain the global best position, which is the global optimal solution.

[0012] This invention establishes a vehicle barrier passage objective function, uses a multi-strategy fusion Genghis Khan shark optimization algorithm to solve the vehicle barrier passage objective function, models the optimization problem, and facilitates algorithm optimization processing. Compared with traditional algorithms, the multi-strategy fusion Genghis Khan shark optimization algorithm avoids falling into local optimal solutions during the iteration process, accelerates convergence speed, improves the population quality and exploration ability of the iteration process, shortens vehicle congestion time, and effectively improves the vehicle barrier passage efficiency.

[0013] Preferably, S4 comprises the following steps: S41, the vehicle gate passage objective function obtains the global optimal solution, obtains the minimum value of the vehicle gate passage average delay time function and the minimum value of the vehicle gate passage queue length function, generates a vehicle gate opening time scheduling plan and a vehicle gate passage quantity scheduling plan, the vehicle gate controls the vehicle gate opening according to the vehicle gate opening time scheduling plan, and the vehicle gate controls the vehicle gate passage quantity according to the vehicle gate passage quantity scheduling plan, and obtains the vehicle gate passage scheduling plan; S42. According to the infrequently used vehicle barrier gate passage data set, the vehicle entry and exit time and vehicle type information of the infrequently used vehicles are obtained, and the priority is set according to the vehicle type information to obtain high-priority vehicles and low-priority vehicles. When a high-priority vehicle passes through the vehicle barrier gate, the number of vehicle barrier gate passages is increased or the vehicle barrier gate opening time is extended, and the vehicle barrier gate passage scheduling plan is adjusted in real time to obtain a final vehicle barrier gate passage scheduling plan.

[0014] The invention generates a vehicle gate passage scheduling plan, adjusts the vehicle gate passage scheduling plan in real time, takes special vehicles and emergencies into consideration, reduces vehicle accidents, and improves the reliability of the scheduling plan.

[0015] This embodiment also discloses a system for intelligent scheduling of vehicle gate traffic based on big data, which specifically includes: a traffic data division and processing module, a traffic flow and abnormality prediction module, an objective function solving module and a traffic scheduling solution generation module; The traffic data division and processing module is used to divide the initial vehicle gate traffic data according to whether it is frequently used and perform pre-processing; The traffic flow and abnormality prediction module is used to predict the traffic flow of vehicle gates and abnormalities of vehicle gates; The objective function solving module is used to solve the vehicle gate passage objective function using the Genghis Khan Shark Optimization Algorithm with multi-strategy fusion; The traffic scheduling scheme generating module is used to generate and adjust the vehicle gate traffic scheduling scheme.

[0016] The present invention has the following beneficial effects: 1. The present invention obtains the commonly used vehicle gate passage data and the infrequently used vehicle gate passage data by identifying, and only processes the commonly used data through data division, thereby reducing the time spent on data processing, and pre-processes the commonly used vehicle gate passage data to improve data validity and accuracy and reduce subsequent prediction errors.

[0017] 2. The invention predicts the vehicle gate traffic flow and vehicle gate anomaly based on neural networks and time series respectively. Since there are fewer vehicle gate anomalies and they do not have good predictability, it is easy to find the change pattern based on time series and accumulated flow to improve the prediction effect; the LSTM recurrent neural network overcomes the gradient vanishing and gradient exploding problems, can determine the time lag length, and has good prediction performance.

[0018] 3. The invention establishes a vehicle barrier passage objective function, uses a multi-strategy fusion Genghis Khan shark optimization algorithm to solve the vehicle barrier passage objective function, models the optimization problem, and facilitates algorithm optimization processing. Compared with traditional algorithms, the multi-strategy fusion Genghis Khan shark optimization algorithm avoids falling into local optimal solutions during the iteration process, accelerates convergence speed, improves the population quality and exploration ability of the iteration process, shortens vehicle congestion time, and effectively improves the vehicle barrier passage efficiency.

[0019] 4. The invention generates a vehicle gate passage scheduling plan, adjusts the vehicle gate passage scheduling plan in real time, takes special vehicles and emergencies into consideration, reduces vehicle accidents, and improves the reliability of the scheduling plan.

[0020] Of course, it is not necessary for any product implementing the present invention to achieve all of the above advantages simultaneously. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0022] Figure 1 FIG. is a schematic flow diagram of the intelligent scheduling of vehicle barrier access provided by the present invention based on big data in vehicle barrier access intelligent scheduling. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0023] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0024] In the description of the present invention, it should be understood that the terms "opening", "upper", "lower", "top", "middle", "inner", etc. indicating the orientation or positional relationship are only for the convenience of describing the invention and simplifying the description, rather than indicating or implying that the components or elements referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention.

[0025] Embodiment 1 Please refer to Figure 1 , this embodiment discloses an intelligent scheduling method for vehicle barrier access based on big data, which specifically includes the following content: S1. Obtain the access records of vehicles entering and leaving the barrier to obtain an initial vehicle barrier access data set, identify and preprocess the common vehicle barrier access data set and the uncommon vehicle barrier access data set in the initial vehicle barrier access data set to obtain a processed vehicle barrier access data set; The S1 includes the following steps: S11. Obtain the access records of vehicles entering and leaving the barrier. The historical access records of vehicles entering and leaving the barrier include vehicle license plate information, vehicle entry and exit times, vehicle type information, etc., which are recorded as vehicle barrier access related information to obtain vehicle barrier access data; using the vehicle license plate information in the vehicle barrier access data as the main body, construct an initial vehicle barrier access data set , where represents the ma subset of vehicle gate passing data, where the subset of vehicle gate passing data is vehicle license plate information, and other vehicle gate passing data is included under each vehicle license plate information; S12. Obtain the vehicle in-out time in the initial vehicle gate passing data set. Within the time period t, count the number of vehicle in-out times, and calculate the vehicle gate passing frequency based on the time period t and the number of vehicle in-out times. Set a frequency threshold. When the vehicle gate passing frequency is greater than the frequency threshold, record the corresponding vehicle license plate information as frequently used vehicles to obtain a frequently used vehicle gate passing data set; otherwise, record the corresponding vehicle license plate information as infrequently used vehicles to obtain an infrequently used vehicle gate passing data set; S13. Preprocess the frequently used vehicle gate passing data set, delete the abnormal vehicle gate passing data in the frequently used vehicle gate passing data set, and perform filling to convert the frequently used vehicle gate passing data set into a processed vehicle gate passing data set. The specific steps are as follows: S131. Within the time period t, divide the time period t by month to obtain a vehicle gate passing month set , where represents the n th vehicle gate passing month; Calculate the average value and variance of the vehicle gate passing data in each vehicle gate passing month according to the vehicle gate passing month set of the initial vehicle gate passing data set, and delete the vehicle gate passing data corresponding to the maximum variance and the minimum variance. Set an average threshold. When the difference between the average value of the vehicle gate passing data and the vehicle gate passing data is less than or equal to the average threshold, the corresponding vehicle gate passing data is recorded as normal vehicle gate passing data; otherwise, obtain abnormal vehicle gate passing data, delete the abnormal vehicle gate passing data, and obtain a processed vehicle gate passing data set; S132. Select any vehicle gate passing month in the vehicle gate passing month set, denoted as the vehicle gate passing month , and then select the adjacent vehicle gate passing month of the vehicle gate passing month . In the processed vehicle gate passing data set, count the missing vehicle gate passing data in the vehicle gate passing month , calculate the average value of the vehicle gate passing data in the adjacent vehicle gate passing month, and use the average value of the vehicle gate passing data to fill the missing vehicle gate passing data in the vehicle gate passing month . Fill the missing vehicle gate passing data in other vehicle gate passing months in turn to obtain a processed vehicle gate passing data set; S2. Based on the processed vehicle gate passing data set, predict the vehicle gate passing flow and vehicle gate anomalies respectively based on neural network and time series, and obtain the vehicle gate passing flow prediction value set and the vehicle gate anomaly prediction value set; The S2 includes the following steps: S21. Obtain the vehicle gate operation information from the processed vehicle gate passing data set. The vehicle gate operation information includes normal operation of the vehicle gate and vehicle gate failure. Divide the processed vehicle gate passing data set into a vehicle gate passing normal operation data set and a vehicle gate passing failure data set according to the vehicle gate operation information; S22. For the vehicle gate passing normal operation data set, train an LSTM recurrent neural network prediction model to predict the vehicle gate passing flow and obtain the vehicle gate passing flow prediction value set. The specific steps are as follows: S221. Obtain the vehicle entry and exit times and the number of vehicle entries and exits in the vehicle gate passing normal operation data set. Use the vehicle entry and exit times as the time series to obtain the vehicle gate passing time series , where represents the th vehicle gate passing time point; According to the number of vehicle entries and exits, calculate the vehicle gate passing flow between each vehicle gate passing time point in turn to obtain the vehicle gate passing flow set. Each vehicle gate passing flow in the vehicle gate passing flow set corresponds to a vehicle gate passing time point; S222. Obtain the historical passing records of vehicles entering and exiting the gate again to obtain the vehicle gate passing historical data set. Preprocess the vehicle gate passing historical data set and calculate the vehicle gate passing historical flow to obtain the vehicle gate passing historical time series and the vehicle gate passing flow sample set. Divide the vehicle gate passing flow sample set into a sample training set and a sample test set; Set the LSTM recurrent neural network to learn according to the stochastic gradient descent method, set the time sliding window, place the time sliding window on the vehicle gate passing historical time series, and obtain the sample training set included in the vehicle gate passing historical time series within the time sliding window. Input it into the LSTM recurrent neural network in turn and iterate until the LSTM recurrent neural network converges to obtain the trained LSTM recurrent neural network; S223. Place the time sliding window on the vehicle gate passing historical time series to obtain the sample test set included in the vehicle gate passing historical time series within the time sliding window. Input it into the trained LSTM recurrent neural network in turn. Set the accuracy threshold. When the accuracy of the output prediction value is greater than the accuracy threshold, obtain the LSTM recurrent neural network prediction model; otherwise, adjust the weights until the accuracy of the output prediction value is greater than the accuracy threshold. S224, according to the vehicle gate passage time series, input the vehicle gate passage flow set into the LSTM recurrent neural network prediction model, and output the vehicle gate passage flow prediction values ​​in sequence to form a vehicle gate passage flow prediction value set; S23, for the vehicle barrier passage failure data set, set the vehicle barrier failure time, take the vehicle barrier failure time as the time series, and obtain the vehicle barrier passage failure time series ,in Indicates vehicle barrier failure time; calculate the vehicle barrier flow rate between each vehicle barrier failure time, and accumulate it according to the vehicle barrier failure time series to obtain the vehicle barrier cumulative flow rate; use the vehicle barrier failure time series as the horizontal coordinate and the vehicle barrier cumulative flow rate as the vertical coordinate to generate a cumulative flow graph, select the curve segment corresponding to the minimum curve slope in the cumulative flow graph, record it as a sample curve segment, determine the smoothing coefficient, use the least squares method to fit the sample curve segment, obtain the vehicle barrier failure time prediction value, record the vehicle barrier failure time prediction value as the vehicle barrier abnormality prediction value, and form a vehicle barrier abnormality prediction value set; S3. According to the vehicle barrier flow prediction value set and the vehicle barrier anomaly prediction set, a vehicle barrier passage objective function is established based on the principle of optimal barrier passage capacity, and the vehicle barrier passage objective function is solved using the Genghis Khan Shark Optimization Algorithm with multi-strategy fusion to obtain a global optimal solution; The S3 comprises the following steps: S31, set the number of vehicles passing through the gate to , combining the vehicle gate flow prediction value set and the vehicle gate abnormality prediction set, based on the principle of optimal gate capacity, determine the average vehicle speed, and obtain the vehicle gate average delay time function and the vehicle gate queue length function; select the minimum value of the vehicle gate average delay time function and the minimum value of the vehicle gate queue length function as the vehicle gate passage objective function, and introduce constraints to meet the vehicle gate opening time is less than or equal to the vehicle gate maximum opening time; S32, introducing the Genghis Khan shark optimization algorithm, and integrating the inferior individual classification strategy and the quasi-adversarial learning strategy to improve the Genghis Khan shark optimization algorithm, to obtain a multi-strategy fusion Genghis Khan shark optimization algorithm, using the multi-strategy fusion Genghis Khan shark optimization algorithm to solve the vehicle gate passage objective function, to obtain the global optimal solution, the specific steps are as follows: S321, the process of solving the objective function of vehicle barrier passage is regarded as a search space, in which there is a Genghis Khan shark population, and the Genghis Khan shark individuals in the Genghis Khan shark population represent candidate solutions of the objective function of vehicle barrier passage; the number of Genghis Khan shark population is set to i , the Genghis Khan shark individual dimension is j , set the current number of iterations to b , initialize the Genghis Khan shark population, and get b +1 iteration i Genghis Khan shark individual location ; During the migration phase, the Genghis Khan shark population updates its position based on the concentration of prey odor, and the odor intensity coefficient is set as , c represents a non-negative constant, represents a random number between the interval [0, 1], then the olfactory intensity , set the b The best Genghis Khan shark individual position at the iteration is , No. b The iteration i Genghis Khan shark individual location , get the b +1 iteration i New location of Genghis Khan shark , at this time, the position To update, ; Get the fitness function value corresponding to the position of the Genghis Khan shark at this time, and record the current best fitness function value and the current worst fitness function value as and , No. i The fitness function value corresponding to the Genghis Khan shark individual is recorded as , introduce the inferior individual classification strategy, calculate pheromone , when the pheromone is less than or equal to 0.3, select the Genghis Khan shark individual location and Genghis Khan shark individual location , for position Update, set Represents a random number of 0 or 1. ; S322, Genghis Khan shark population is in the foraging stage, and the step length parameter is set to , based on the best Genghis Khan shark individual position About Location Update again, this time introduce the quasi-adversarial learning strategy, set the upper and lower bounds of the search space to be and , dual value , dual position , calculate the fitness function value corresponding to the dual position, compare it with the current best fitness function, and use the dual position to replace the position , to achieve the position Updates; In the self-protection stage of the Genghis Khan shark population, the Genghis Khan shark individuals use self-protection strategies to position Update to get b +1 iteration i The final position of the Genghis Khan shark individual is obtained, the Genghis Khan shark population is screened, and the next iteration is entered; the maximum number of iterations is set, and when the current number of iterations reaches the maximum number of iterations, the iteration is stopped to obtain the global best position, which is the global optimal solution; S4, generating a vehicle gate passage scheduling plan according to the global optimal solution, and introducing the uncommon vehicle gate passage data set to adjust the vehicle gate passage scheduling plan in real time to obtain a final vehicle gate passage scheduling plan; The S4 comprises the following steps: S41, the vehicle gate passage objective function obtains the global optimal solution, obtains the minimum value of the vehicle gate passage average delay time function and the minimum value of the vehicle gate passage queue length function, generates a vehicle gate opening time scheduling plan and a vehicle gate passage quantity scheduling plan, the vehicle gate controls the vehicle gate opening according to the vehicle gate opening time scheduling plan, and the vehicle gate controls the vehicle gate passage quantity according to the vehicle gate passage quantity scheduling plan, and obtains the vehicle gate passage scheduling plan; S42. According to the infrequently used vehicle barrier gate passage data set, the vehicle entry and exit time and vehicle type information of the infrequently used vehicles are obtained, and the priority is set according to the vehicle type information to obtain high-priority vehicles and low-priority vehicles. When a high-priority vehicle passes through the vehicle barrier gate, the number of vehicle barrier gate passages is increased or the vehicle barrier gate opening time is extended, and the vehicle barrier gate passage scheduling plan is adjusted in real time to obtain a final vehicle barrier gate passage scheduling plan.

[0026] Example 2 This embodiment also discloses a system for intelligent scheduling of vehicle gate traffic based on big data, which specifically includes: a traffic data division and processing module, a traffic flow and abnormality prediction module, an objective function solving module and a traffic scheduling solution generation module; The traffic data division and processing module is used to divide the initial vehicle gate traffic data according to whether it is frequently used and perform pre-processing; The traffic flow and anomaly prediction module is used to predict the vehicle barrier traffic flow and vehicle barrier anomalies; The objective function solving module is used to solve the vehicle barrier traffic objective function using a multi-strategy fusion Genghis Khan shark optimization algorithm; The traffic scheduling plan generation module is used to generate and adjust the vehicle barrier traffic scheduling plan.

[0027] In the description of this specification, the description referring to terms such as "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0028] The preferred embodiments of the invention disclosed above are only used to help explain the invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and variations can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principle and practical application of the invention, so that those skilled in the art can well understand and utilize the invention.

Claims

1. The intelligent dispatching method for vehicle gate passage based on big data is characterized by: The steps include: S1. Obtain the passage records of vehicles entering and exiting the gate to obtain an initial vehicle gate passage data set, identify and pre-process the commonly used vehicle gate passage data set and the uncommon vehicle gate passage data set in the initial vehicle gate passage data set, and obtain a processed vehicle gate passage data set; S2. According to the processed vehicle barrier passage data set, the vehicle barrier passage flow and vehicle barrier anomaly are predicted based on the neural network and the time series, respectively, to obtain a vehicle barrier passage flow prediction value set and a vehicle barrier anomaly prediction value set; S3, according to the vehicle gate flow prediction value set and the vehicle gate abnormality prediction set, based on the principle of optimal gate capacity, establish a vehicle gate passage objective function, solve the vehicle gate passage objective function, and obtain a global optimal solution; S4. Generate a vehicle gate traffic scheduling plan based on the global optimal solution, and introduce the infrequently used vehicle gate traffic data set to adjust the vehicle gate traffic scheduling plan in real time to obtain a final vehicle gate traffic scheduling plan.

2. The method for intelligent dispatching of vehicle gate passage based on big data according to claim 1 is characterized in that: The S1 comprises the following steps: S11, obtaining the passage records of vehicles entering and exiting the gate, obtaining the vehicle gate passage data, and forming an initial vehicle gate passage data set; S12, calculating the frequency of vehicles entering and exiting the gate, and dividing the initial vehicle gate passage data set into a frequently used vehicle gate passage data set and an infrequently used vehicle gate passage data set; S13, pre-processing the commonly used vehicle barrier gate passage data set, deleting abnormal vehicle barrier gate passage data in the commonly used vehicle barrier gate passage data set, and performing data filling to obtain a processed vehicle barrier gate passage data set.

3. The method for intelligent dispatching of vehicle gate passage based on big data according to claim 2 is characterized in that: The S13 comprises the following steps: S131. Within the time period t, an average threshold is set, and the average value of the vehicle gate passage data and the average threshold are compared, and abnormal vehicle gate passage data are identified and deleted to obtain a processed vehicle gate passage data set; S132. Divide the time period t according to months, and use the mean of vehicle barrier passage data in adjacent vehicle barrier passage months to fill in the missing data, so as to obtain a processed vehicle barrier passage data set.

4. The method for intelligent dispatching of vehicle gate passage based on big data according to claim 3 is characterized in that: The S2 comprises the following steps: S21, dividing the processed vehicle barrier gate passage data set into a vehicle barrier gate passage normal operation data set and a vehicle barrier gate passage fault data set; S22, training the normal operation data set of the vehicle gate to obtain an LSTM recurrent neural network prediction model, predicting the vehicle gate flow rate, and obtaining a vehicle gate flow rate prediction value set; S23. Obtain a vehicle barrier gate passage failure time series from the vehicle barrier gate passage failure data set, predict vehicle barrier gate anomalies, and obtain a vehicle barrier gate anomaly prediction value set.

5. The method for intelligent dispatching of vehicle gate passage based on big data according to claim 4 is characterized in that: The S22 comprises the following steps: S221, obtaining a vehicle gate passage time series and a vehicle gate passage flow set according to the vehicle gate passage normal operation data set; S222, obtaining the historical passage records of vehicles entering and exiting the gate, obtaining a set of historical data on the passage of the vehicle gate, preprocessing the set of historical data on the passage of the vehicle gate and calculating the historical flow of the vehicle gate, obtaining a historical time series of the passage of the vehicle gate and a sample set of the passage flow of the vehicle gate, and dividing the sample set of the passage flow of the vehicle gate into a sample training set and a sample test set; setting a time sliding window, the LSTM recurrent neural network learns according to the stochastic gradient descent method, placing the time sliding window in the historical time series of the passage of the vehicle gate, obtaining a sample training set contained in the historical time series of the passage of the vehicle gate in the time sliding window, and sequentially inputting them into the LSTM recurrent neural network, iterating until the LSTM recurrent neural network converges, and obtaining a trained LSTM recurrent neural network; S223, placing the time sliding window in the vehicle gate passage history time series, obtaining the sample test set contained in the vehicle gate passage history time series in the time sliding window, inputting them into the trained LSTM recurrent neural network in sequence, setting the accuracy threshold, and obtaining the LSTM recurrent neural network prediction model when the output prediction value accuracy is greater than the accuracy threshold, otherwise adjusting the weight until the output prediction value accuracy is greater than the accuracy threshold; S224. According to the vehicle barrier passage time series, the vehicle barrier passage flow set is input into the LSTM recurrent neural network prediction model, and the vehicle barrier passage flow prediction values ​​are output in sequence to form a vehicle barrier passage flow prediction value set.

6. The method for intelligent dispatching of vehicle gate passage based on big data according to claim 5 is characterized in that: The S3 comprises the following steps: S31, set the number of vehicles passing through the gate to , combining the vehicle gate flow prediction value set and the vehicle gate abnormality prediction set, based on the principle of optimal gate capacity, determining the average vehicle speed, and obtaining the vehicle gate average delay time function and the vehicle gate queue length function; The minimum value of the average delay time function of the vehicle gate and the minimum value of the queue length function of the vehicle gate are selected as the vehicle gate passage objective function, and the constraint condition is introduced to meet the requirement that the vehicle gate opening time is less than or equal to the vehicle gate maximum opening time; S32. Use the inferior individual classification strategy and the quasi-adversarial learning strategy to improve the Genghis Khan shark optimization algorithm to obtain a multi-strategy fusion Genghis Khan shark optimization algorithm, and use the multi-strategy fusion Genghis Khan shark optimization algorithm to solve the vehicle gate passage objective function to obtain the global optimal solution.

7. The method for intelligent dispatching of vehicle gate passage based on big data according to claim 6 is characterized in that: The S32 comprises the following steps: S321, regarding the process of solving the objective function of the vehicle barrier passage as a search space, wherein there is a Genghis Khan shark population in the search space, and the Genghis Khan shark individuals in the Genghis Khan shark population represent candidate solutions of the objective function of the vehicle barrier passage; Set the Genghis Khan shark population to i , the Genghis Khan shark individual dimension is j , the current number of iterations is b , initialize the Genghis Khan shark population, and get b +1 iteration i Genghis Khan shark individual location ; During the migration phase, the Genghis Khan shark population updates its position based on the concentration of prey odor, and the odor intensity coefficient is set as , c represents a non-negative constant, represents a random number between the interval [0, 1], then the olfactory intensity , set the b The best Genghis Khan shark individual position at the iteration is , for position Make updates; Get the fitness function value corresponding to the position of the Genghis Khan shark at this time, and record the current best fitness function value and the current worst fitness function value as and , No. i The fitness function value corresponding to the Genghis Khan shark individual is recorded as , introduce the inferior individual classification strategy, calculate pheromone , when the pheromone is less than or equal to 0.3, select Genghis Khan shark individual location and Genghis Khan shark individual location , for position Update, set Represents a random number of 0 or 1. ; S322, Genghis Khan shark population is in the foraging stage, and the step length parameter is set to , based on the best Genghis Khan shark individual position About Location Update again, this time introduce the quasi-adversarial learning strategy, set the upper and lower bounds of the search space to be and , dual value , dual position , calculate the fitness function value corresponding to the dual position, compare it with the current best fitness function, and use the dual position to replace the position , to achieve the position Updates; In the self-protection stage of the Genghis Khan shark population, the Genghis Khan shark individuals use self-protection strategies to position Update to get b +1 iteration i The final position of the Genghis Khan shark individual is obtained, and the Genghis Khan shark population is screened to enter the next iteration; The maximum number of iterations is set, and when the current number of iterations reaches the maximum number of iterations, the iteration is stopped to obtain the global best position, which is the global optimal solution.

8. The method for intelligent dispatching of vehicle gate passage based on big data according to claim 7 is characterized in that: The S4 comprises the following steps: S41, the vehicle gate passage objective function obtains a global optimal solution, generates a vehicle gate opening time scheduling plan and a vehicle gate passage quantity scheduling plan, and obtains a vehicle gate passage scheduling plan; S42. According to the infrequently used vehicle gate passage data set, the vehicle gate passage scheduling plan is adjusted in real time to obtain a final vehicle gate passage scheduling plan.

9. A system for implementing the vehicle gate passage intelligent scheduling method based on big data as described in any one of claims 1 to 8, characterized in that: Specifically include: Traffic data division and processing module, traffic flow and abnormality prediction module, objective function solution module and traffic scheduling plan generation module; The traffic data division and processing module is used to divide and pre-process the initial vehicle gate traffic data according to whether it is frequently used; The traffic flow and abnormality prediction module is used to predict the traffic flow of vehicle gates and abnormalities of vehicle gates; The objective function solving module is used to solve the vehicle gate passage objective function using the Genghis Khan Shark Optimization Algorithm with multi-strategy fusion; The traffic scheduling scheme generating module is used to generate and adjust the vehicle gate traffic scheduling scheme.

Images