Operation management control method for networked vehicle

Abstract

The invention discloses an operation management control method for networked vehicles, and the method comprises the steps: obtaining the historical traffic flow of a to-be-predicted scenic region, determining the predicted traffic flow of the scenic region based on an ARIMA model of the scenic region and the historical traffic flow, and enabling the ARIMA model to convert a non-stationary time sequence into a stationary time sequence, and carrying out the prediction, thereby achieving the prediction of the vehicle flow of the scenic region. The method can predict the predicted traffic flow when the scenic spot is not affected, and determines the influence weight of holidays and festivals on the traffic flow in the prediction time period based on the number of holidays and festivals in the prediction time period. And according to the predicted traffic flow and the influence weight, determining the target traffic flow of the scenic region in the prediction time period, thereby referring to the conditions of holidays and festivals when determining the traffic flow of the scenic region, and effectively improving the accuracy of predicting the traffic flow of the scenic region.

Description

Technical Field

[0001] The present application relates to the field of connected vehicle control technology, and in particular to an operation management and control method for a connected vehicle. Background Art

[0002] The product positioning and overall planning of scenic spot projects have a profound impact on the development of the tourism industry, and the prediction of vehicle flow is an important part of the formulation of tourism development plans, and is important data that scenic spot managers need to understand in their daily management work. Scenic spot managers estimate the number of scenic spot vehicles by analyzing the trend of tourism demand, providing a reference for the arrangement of scenic spot flow control, traffic diversion and safety management. Summary of the invention

[0003] In an exemplary embodiment of the present application, a method for operation management and control of a connected vehicle is provided to solve the problem of difficult and large error in vehicle flow prediction in related technologies.

[0004] The present application provides an operation management control method for a connected vehicle, which includes: Receive a prediction request, and obtain an identifier of a scenic spot to be predicted and a prediction time period carried in the prediction request; Based on the autoregressive integrated moving average ARIMA model pre-constructed for the identified scenic spot and the traffic volume of each historical time period of the scenic spot, predicting the predicted traffic volume of the scenic spot in the predicted time period; Determine the number of days that include holidays in the forecast time period, obtain random numbers that are the same as the number of days, and determine the impact weight of holidays on the traffic flow in the forecast time period according to the sum of each random number obtained; wherein the greater the number of days, the greater the impact weight; The target traffic flow of the scenic spot in the predicted time period is determined according to the predicted traffic flow and the impact weight.

[0005] Specifically, the random numbers having the same number as the number of days include: According to the number of days of the holiday, determine the index value corresponding to the number of days; In a normal distribution whose mean is a preset value and whose standard deviation is the indicator value, determine a number of random numbers that is the same as the number of days.

[0006] Specifically, determining the index value corresponding to the number of holidays according to the number of holidays includes: Obtain the maximum and minimum number of holidays in the year to which the forecast time period belongs, and determine the difference between the maximum and minimum number of holidays; According to the ratio of the number of days containing holidays in the predicted time period to the difference, the index value corresponding to the number of days is determined.

[0007] Specifically, after determining the number of days including holidays in the prediction time period and before obtaining the same number of random numbers as the number of days, the method further includes: Obtain the number of days affected by the time interval between the predicted time period and holidays within a preset time interval, obtain the sum of the affected days and the days, and use the sum to adjust the days.

[0008] Specifically, the method further includes: Obtain each historical ticket price for each historical time period saved for the scenic spot, and the ticket price for the predicted time period, and obtain each historical weather for each historical time period saved for the scenic spot; Input each historical ticket price, the ticket price, each historical weather and the traffic volume in each historical time period into a pre-trained traffic volume prediction model, and obtain the residual traffic volume affected by the ticket price and weather output by the traffic volume prediction model; The target vehicle flow rate is adjusted based on the residual vehicle flow rate.

[0009] Specifically, the traffic flow prediction model is trained in the following way: Obtain any data set in the training set, wherein the data set includes ticket prices for multiple historical time periods, ticket prices for predicted time periods, weather for the multiple historical time periods, and vehicle flows for the multiple historical time periods, and obtain the labeled vehicle flows for the predicted time periods affected by weather and ticket prices saved for the data set; Inputting the data group into an original vehicle flow prediction model to obtain an output vehicle flow output by the original vehicle flow prediction model; The original vehicle flow prediction model is trained based on the labeled vehicle flow and the output vehicle flow.

[0010] Specifically, the method further includes: Acquire the Fourier coefficients saved for the scenic spot, determine the Fourier function corresponding to the Fourier coefficients; and determine the target period corresponding to the predicted time period; Based on the Fourier function and the target period, determining seasonal traffic flow affected by seasons corresponding to the prediction time period; The target traffic volume is adjusted based on the seasonal traffic volume.

[0011] The embodiments of the present application have the following beneficial effects: the historical traffic flow of the scenic spot to be predicted is obtained, and the predicted traffic flow of the scenic spot is determined based on the ARIMA model and the historical traffic flow of the scenic spot. Since the ARIMA model can convert a non-stationary time series into a stationary time series and then perform prediction, the predicted traffic flow when the scenic spot is not affected can be predicted, and based on the number of holidays in the prediction time period, the impact weight of holidays on the traffic flow in the prediction time period is determined; according to the predicted traffic flow and the impact weight, the target traffic flow of the scenic spot in the prediction time period is determined, so that when determining the traffic flow of the scenic spot, the situation of holidays is referred to, thereby effectively improving the accuracy of the traffic flow prediction of the scenic spot. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0013] Figure 1 A schematic diagram of a flow chart of an operation management control method for a networked vehicle provided in an embodiment of the present application is exemplarily shown; Figure 2 A schematic diagram of a process for determining an influence weight in an operation management and control method for a networked vehicle provided in an embodiment of the present application is exemplarily shown; Figure 3 A schematic diagram of a process for determining residual vehicle flow in an operation management and control method for a networked vehicle provided in an embodiment of the present application is exemplarily shown; Figure 4 A flow chart of determining seasonal vehicle flow in an operation management and control method for a connected vehicle provided in an embodiment of the present application is exemplified. DETAILED DESCRIPTION

[0014] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0015] To further illustrate the technical solution provided by the embodiment of the present application, this is described in detail below in conjunction with the accompanying drawings and specific implementation methods. Although the embodiment of the present application provides the method operation steps shown in the following embodiments or drawings, more or fewer operation steps may be included in the method based on routine or no creative labor. In the steps where there is no necessary causal relationship logically, the execution order of these steps is not limited to the execution order provided in the embodiment of the present application.

[0016] refer to Figure 1 As shown, the present application provides an operation management control method for a connected vehicle, which includes: S101: receiving a prediction request, and obtaining an identifier of a scenic spot to be predicted and a prediction time period carried in the prediction request.

[0017] The vehicle flow prediction method provided in the embodiment of the present application is applied to an electronic device, which may be a smart device such as a PC or a server.

[0018] In order to accurately and effectively predict traffic flow, the electronic device can first receive a prediction request, wherein the prediction request can be sent by the user to the electronic device based on the device used by the user. The user can select the identification of the scenic area to be predicted and the prediction time period on the preset page of the device used by the user, wherein the identification of the scenic area to be predicted can be the name of the scenic area to be predicted. After the user clicks the preset button, the device used by the user sends a prediction request to the electronic device. After receiving the prediction request, the electronic device can obtain the identification of the scenic area to be predicted and the prediction time period carried in the prediction request.

[0019] S102: Based on the ARIMA model pre-constructed for the identified scenic spot and the traffic volume of each historical time period of the scenic spot, predict the predicted traffic volume of the scenic spot in the predicted time period.

[0020] Historical traffic flow can reflect the behavior patterns and preferences of tourist vehicles. By analyzing historical traffic flow data, we can understand the changing trend of the traffic flow of scenic spots in different time periods. Therefore, after obtaining the identification of the scenic spot to be predicted and the predicted time period, the electronic device can obtain the ARIMA model constructed for the identified scenic spot and obtain the traffic flow of each historical time period of the scenic spot.

[0021] The electronic device can predict the predicted traffic flow of the scenic spot in the predicted time period based on the ARIMA model pre-constructed for the identified scenic spot and each acquired traffic flow.

[0022] It should be noted that the ARIMA model is used for self-prediction of a single variable time series. AR is autoregression, MA is moving average, I is the difference operation, and d is the number of differences required for the time series to become stationary. The ARIMA model refers to a model that converts a non-stationary time series into a stationary time series, and then regresses the dependent variable only on its lagged value and the present value and lagged value of the random error term. The modeling process based on the ARIMA time series model in advance is as follows: First, obtain the sequence of traffic flow in the scenic area for a period of time, that is, ,in , and test the stationarity of the sequence. Stationarity can be divided into weak stationary processes and strict stationary processes due to its own characteristics. Usually, the basic assumption in econometrics is to require the economic variable process under study to have weak stationarity. Therefore, it can be determined whether the sequence is stationary. If the sequence is a non-stationary sequence, the sequence is differentially operated until the stable d-order differential data is obtained. If it is a stationary sequence, a white noise test is performed, usually using the Ljung-Box test. If it is a white noise sequence, it cannot be predicted and the analysis ends. If it is a non-white noise sequence, the time series data is fitted with a model, including model identification, parameter estimation, model verification, residual sequence test and other steps, and finally the fitted ARIMA model is obtained. And establish ARIMA (p, d, q) models with different p and q orders, and compare the AIC values. Select the p and q with the minimum AIC, and the p and q based on the minimum AIC and the d when reaching stability are used as the parameters of the final ARIMA model.

[0023] In a possible implementation, the electronic device may verify whether the sequence is stable in the following manner: , μ is a constant; and, ,for If the above formula is satisfied, then it is called is a weakly stationary process. The covariance of a weakly stationary process is constant and only depends on the interval k between the two random variables in the process, but has nothing to do with the starting position t. The covariance here is It is also called the autocovariance function.

[0024] In a possible implementation, the electronic device may also use ADF to test the stationarity of the sequence. The specific model of the ADF test is shown in the following formula: ; Among them, α is the autoregressive coefficient, is the sequence of traffic flow, is a random error term.

[0025] On this basis, the null hypothesis and alternative hypothesis are proposed: ; ; when H When 0 holds, it means that the tested sequence is non-stationary. H When 1 holds, it means that the verified sequence is stationary.

[0026] Specifically, how to construct the ARIMA model is an existing technology and will not be described in detail here.

[0027] The formula for the Ljung-Box test is usually: ; Among them, Q approximately obeys the chi-square distribution with Kpq degrees of freedom. By judging whether Q exceeds the critical value of the chi-square distribution, it is judged whether there is correlation within the lag order K, that is, if there is correlation, it is a non-white noise sequence, and if there is no correlation, it is a white noise sequence. Specifically, how to perform the Ljung-Box test is an existing technology and will not be repeated here.

[0028] The final fitted ARIMA model is as follows: Autoregressive prediction: The formula definition of the P-order autoregressive process is: ,in is the current value, μ is the constant term, p is the order, is the autocorrelation coefficient, It's an error.

[0029] Moving average model: Q-order autoregressive process formula definition: , is the correlation coefficient of the demand solution.

[0030] Therefore, the autoregressive moving average model can be obtained ; in, is the current value, μ is the constant term, p and q are the orders, specifically p is the autoregressive term, q is the moving average term, is the autocorrelation coefficient, It's an error.

[0031] S103: Determine the number of days that include holidays in the predicted time period, obtain random numbers that are the same as the number of days, and determine the impact weight of holidays on the traffic volume in the predicted time period based on the sum of each random number obtained; wherein, the greater the number of days, the greater the impact weight.

[0032] Since statutory holidays also have a high degree of impact on the traffic volume of scenic spots, the electronic device can determine the impact weight of holidays. The electronic device can determine the number of days containing holidays in the prediction time period, obtain the random number, determine the sum of each obtained random number, and determine the impact weight of holidays on the traffic volume in the prediction time period according to the sum.

[0033] It should be noted that, generally, the greater the number of days, the greater the impact weight determined.

[0034] S104: Determine the target traffic flow of the scenic spot in the predicted time period according to the predicted traffic flow and the impact weight.

[0035] After determining the predicted traffic volume and the influence weight, the electronic device can determine the target traffic volume of the scenic spot in the predicted time period according to the preset traffic volume and the influence weight. In a possible implementation, the electronic device can determine the product of the predicted traffic volume and the influence weight, and determine the product as the target traffic volume of the scenic spot in the predicted time period.

[0036] The embodiment of the present application is a time series prediction method applied to the direction of regional traffic flow prediction, which can effectively solve the problems of the traffic flow prediction method in the prior art, such as lack of data support, incomplete consideration of factors, and poor adaptability, and has innovative and practical value.

[0037] In the embodiment of the present application, the electronic device obtains the historical traffic flow of the scenic area to be predicted, and determines the predicted traffic flow of the scenic area based on the ARIMA model and the historical traffic flow of the scenic area. Since the ARIMA model can convert non-stationary time series into stationary time series and then make predictions, the predicted traffic flow when the scenic area is not affected can be predicted, and based on the number of holidays in the prediction time period, the impact weight of holidays on the traffic flow in the prediction time period is determined; according to the predicted traffic flow and the impact weight, the target traffic flow of the scenic area in the prediction time period is determined, so that when determining the traffic flow of the scenic area, the situation of holidays is referred to, which effectively improves the accuracy of the traffic flow prediction of the scenic area.

[0038] In order to determine the impact of holidays on traffic flow, based on the above embodiment, in the embodiment of the present application, the random numbers having the same number as the days include: According to the number of days of the holiday, determine the index value corresponding to the number of days; In a normal distribution whose mean is a preset value and whose standard deviation is the indicator value, determine a number of random numbers that is the same as the number of days.

[0039] In order to determine the impact of holidays on traffic flow, the electronic device can determine the index value corresponding to the number of holidays. In a possible implementation, the electronic device can obtain the index value saved for the number of days, or determine the number of days as the corresponding index. It should be noted that the more days there are, the larger the index value. After obtaining the index value, the electronic device can obtain a normal distribution with a mean value of a preset value and a standard deviation of the index value, and determine the random number in the normal distribution. The preset value can be 0. The normal distribution can be expressed as: , a is the mean, v is the standard deviation, and the normal distribution is affected by the indicator v. In a possible implementation, the default value of the standard deviation may be 10.

[0040] It should be noted that the larger the standard deviation, the greater the impact of holidays on traffic flow; the smaller the standard deviation, the smaller the impact of holidays on traffic flow.

[0041] By adopting the method provided in the embodiment of the present application, the random numbers drawn will follow the characteristics of normal distribution, that is, most of the values ​​will be concentrated near the mean, and the probability of occurrence of values ​​far from the mean will gradually decrease.

[0042] In the real world, there are many holidays besides weekends, and different countries have different holidays. Through the descriptions of holidays in various countries on Wikipedia, we have collected special holidays in various countries. In addition to holidays, staff can also set necessary holidays according to their own circumstances, such as the annual championship game (The Super Bowl), Double Eleven, etc.

[0043] In order to determine the index value corresponding to the number of days, based on the above embodiments, in the embodiment of the present application, the index value corresponding to the number of days according to the number of holidays is determined, including: Obtain the maximum and minimum number of holidays in the year to which the forecast time period belongs, and determine the difference between the maximum and minimum number of holidays; According to the ratio of the number of days containing holidays in the predicted time period to the difference, the index value corresponding to the number of days is determined.

[0044] Since each holiday has a different degree of influence on the time series, for example, the Spring Festival and National Day are seven-day holidays, while Labor Day and other holidays are shorter. Therefore, different holidays can be regarded as independent models, so the index values ​​corresponding to different holidays are also different. In the embodiment of the present application, the electronic device can determine the corresponding index value according to the number of days of the holiday.

[0045] In order to determine the index value corresponding to the number of days, the electronic device may obtain the maximum number of days and the minimum number of days of holidays in the year to which the forecast time period belongs, and determine the difference between the maximum number of days and the minimum number of days. And according to the ratio of the number of days of holidays included in the forecast time period to the difference, the index value corresponding to the number of days is determined. In a possible implementation, the electronic device may determine the ratio as the index value corresponding to the number of days. The electronic device may also determine the product of the ratio and a preset value as the index value corresponding to the number of days.

[0046] Specifically, the electronic device may use the following formula to determine the index value corresponding to the number of days: ; in, is the indicator value corresponding to the number of days, 20 is the preset value, For that number of days, is the maximum number of holidays in the year to which the holiday belongs. The minimum number of holidays in the year to which the holiday belongs.

[0047] This process can also be called hashing by normalizing the maximum and minimum number of days in the holidays.

[0048] Figure 2 A schematic diagram of a process for determining an influence weight provided in an embodiment of the present application, the process includes the following steps: S201: Obtain the maximum and minimum number of holidays in the year to which the forecast time period belongs.

[0049] S202: Determine the difference between the maximum number of days and the minimum number of days.

[0050] S203: Determine the index value corresponding to the number of days containing holidays in the forecast time period according to the ratio of the number of days containing holidays to the difference.

[0051] S204: In a normal distribution whose mean is a preset value and whose standard deviation is an index value, determine a number of random numbers that include holidays and whose number is the same as the number of days in the prediction time period.

[0052] S205: Determine the influence weight of holidays on the traffic flow in the predicted time period according to the sum of each obtained random number.

[0053] In order to improve the accuracy of traffic flow prediction, based on the above embodiments, after determining the number of days including holidays in the prediction time period in the embodiment of the present application, before obtaining the random numbers that are the same as the number of days, the method further includes: Obtain the number of days affected by the time interval between the predicted time period and holidays within a preset time interval, obtain the sum of the affected days and the days, and use the sum to adjust the days.

[0054] In actual scenarios, the traffic volume before and after holidays will also have an impact. Therefore, the electronic device can obtain the number of affected days in the predicted time period and the time interval between the holidays within the preset time interval. The preset time interval can be 1 day or 2 days, etc., and obtain the sum of the affected days and the number of holidays in the predicted time period, use the sum to adjust the number of holidays, and subsequently determine the random value of the adjusted number of days.

[0055] The electronic device can use the following formula to determine the impact weight: ; in, and ,in, To determine the impact weight, is the weight corresponding to the tth random number, is the i-th random number, and L is the number of days after the adjustment.

[0056] In order to accurately and effectively determine the traffic volume of the scenic area, based on the above embodiments, in the embodiment of the present application, the method further includes: Obtain each historical ticket price for each historical time period saved for the scenic spot, and the ticket price for the predicted time period, and obtain each historical weather for each historical time period saved for the scenic spot; Input each historical ticket price, the ticket price, each historical weather and the traffic volume in each historical time period into a pre-trained traffic volume prediction model, and obtain the residual traffic volume affected by the ticket price and weather output by the traffic volume prediction model; The target vehicle flow rate is adjusted based on the residual vehicle flow rate.

[0057] Weather and climate data can directly affect the travel plans of tour buses, especially in outdoor scenic spots. For example, if the weather forecast shows that there will be bad weather in the next few days, the tour bus may choose to cancel or postpone the trip. And the ticket price has a great influence on the decision of the tour bus. If the ticket price increases, the tour bus may choose other scenic spots. In order to improve the accuracy of the determination of the traffic volume of the scenic spot, the electronic device can determine the residual traffic volume affected by the ticket price and weather based on the pre-trained traffic volume prediction model.

[0058] Specifically, the electronic device can obtain each historical ticket price for each historical time period saved for the scenic spot, and the ticket price for the predicted time period. The electronic device also obtains each historical weather for each historical event segment saved for the scenic spot. In order to accurately determine the traffic flow, the electronic device locally stores a pre-trained traffic flow prediction model, wherein the trained traffic flow prediction model is used to predict the impact of external factors such as weather and ticket prices on traffic flow. The electronic device can input each historical ticket price, the ticket price for the predicted time period, each historical weather and the traffic flow for each historical time period into the pre-trained traffic flow prediction model, obtain the residual traffic flow affected by the ticket price and weather output by the traffic flow prediction model, and adjust the target traffic flow based on the residual traffic flow.

[0059] Figure 3 A schematic diagram of a process for determining residual vehicle flow provided in an embodiment of the present application, the process comprising the following steps: In actual scenarios, you can first obtain each historical ticket price, or you can first obtain each historical weather. Figure 3 The following takes the example of obtaining historical ticket prices first.

[0060] S301: Obtain each historical ticket price for each historical time period saved for the scenic spot, and the ticket price for the predicted time period.

[0061] S302: Obtain each historical weather of each historical time period saved for the scenic spot.

[0062] S303: Input each historical ticket price, ticket price, each historical weather and the traffic volume in each historical time period into the pre-trained traffic volume prediction model.

[0063] S304: Obtain the residual traffic flow affected by ticket prices and weather output by the traffic flow prediction model.

[0064] Previous scientific research theories and application confirmations have shown that a single periodic indicator sequence as a simple time series analysis and modeling is difficult and inaccurate, and the prediction and modeling using a single piece of information in isolation is not robust. According to the actual output of various statistical indicators in a complex system, auxiliary information can be analyzed to help predict future index returns, and this information is reflected in the deviation of the observed indicator from its inherent dynamic trend. Its impact is nonlinear and not easy to model using parameter models.

[0065] Easy to understand, a time series forecast can be broken down as follows: ; in, for The linear part of for The nonlinear part.

[0066] Therefore, if For the periodic time series indicators that need to be predicted in this application, then Measure its internal dynamics, and Weigh the impact of recent information. The internal dynamic trend of a general complex system will not change much in a short period of time, so its linear trend part can be the target traffic volume predicted by the embodiment of this application. The nonlinear part is affected by multiple information, and its effect is non-parametric. Since periodic traffic indicators are often affected by holidays, climate and other factors, the prediction framework will combine multiple calculation parts. The nonlinear part can be the residual traffic volume described in the embodiment of this application.

[0067] Therefore, in a possible implementation, the sum of the target vehicle flow and the residual vehicle flow may be determined, and the target vehicle flow may be updated using the sum.

[0068] In actual scenarios, discussions and sharing on social media can provide real-time information about the popularity and attractiveness of scenic spots. By monitoring topics and trends on social media, the interests and needs of potential tour buses can be understood. In addition, marketing activities and promotions can attract more tour buses. Knowing when to conduct marketing activities and promotions, as well as their expected effects, can help predict future traffic volume. And economic indicators will have an impact on traffic volume. If the local economic situation is good and residents' income is stable, then the traffic volume may be higher. On the contrary, if the economic situation is not good and residents' income is declining, then tour buses may reduce travel. In addition, if the scenic area has convenient transportation, then the traffic volume may be higher. On the contrary, traffic congestion or road closures may reduce traffic volume. Therefore, in the embodiment of the present application, topics and trends on social media, scenic area marketing activities and promotions, economic indicators and scenic area traffic can also be input into the traffic volume prediction model, and the residual traffic volume output by the traffic volume prediction model can also include traffic volume affected by topics and trends on social media, scenic area marketing activities and promotions, economic indicators and scenic area traffic.

[0069] In order to accurately predict the traffic flow, based on the above embodiments, in the embodiment of the present application, the traffic flow prediction model is trained in the following manner: Obtain any data set in the training set, wherein the data set includes ticket prices for multiple historical time periods, ticket prices for predicted time periods, weather for the multiple historical time periods, and vehicle flows for the multiple historical time periods, and obtain the labeled vehicle flows for the predicted time periods affected by weather and ticket prices saved for the data set; Inputting the data group into an original vehicle flow prediction model to obtain an output vehicle flow output by the original vehicle flow prediction model; The original vehicle flow prediction model is trained based on the labeled vehicle flow and the output vehicle flow.

[0070] In order to train the traffic flow prediction model, the electronic device can obtain any data group in the training set, wherein the data group includes ticket prices for multiple historical time periods, ticket prices for predicted time periods, weather for multiple historical time periods, and traffic flow for the multiple historical time periods, and obtain the labeled traffic flow affected by weather and ticket prices for the predicted time period saved for the data group.

[0071] The electronic device can input the data group into the original traffic flow prediction model, obtain the output traffic flow output by the original traffic flow prediction model, and train the original traffic flow prediction model based on the labeled traffic flow and the output traffic flow.

[0072] The original vehicle flow prediction model is trained in the above manner, and when the preset conditions are met, the trained original vehicle flow prediction model is obtained. The preset conditions may be that the number of data groups in the training set whose output vehicle flow and labeled vehicle flow are consistent after the original vehicle flow prediction model is trained is greater than a set number; or the number of iterations of training the original vehicle flow prediction model reaches the set maximum number of iterations, etc. Specifically, the embodiments of the present application do not limit this.

[0073] In a possible implementation, when training the traffic flow prediction model, multiple data sets can be input into the original traffic flow prediction model, and predictions are made based on the output traffic flow and the labeled traffic flow corresponding to the multiple data sets. The trained traffic flow prediction model is a typical artificial neural network (ANN), which generally includes input items, hidden layers, and output items. If the input item is , the target item is , the output is , where the target item is the labeled traffic flow and the output item is the output traffic flow. The training of the ANN neural network is to Next, find the weight number , so that the following objective function is minimized: ; in ; An important property of ANN is that by selecting different weights of parameters, its form can fit any function. Then the transformation functions u1(x) and u2(x) can be taken as the same sigmoid function, which has the following form: ; When using ANN, the data is divided into a training set and a test set. A neural network can be trained using the data from the training set and then tested on the test set. However, the ANN neural network consists of three layers including the input layer, hidden layer, and output layer, so it is suitable for showing the nonlinear non-parametric relationship between input information and output information.

[0074] In order to improve the accuracy of vehicle flow prediction, based on the above embodiments, in the embodiment of the present application, the method further includes: Acquire the Fourier coefficients saved for the scenic spot, determine the Fourier function corresponding to the Fourier coefficients; and determine the target period corresponding to the predicted time period; Based on the Fourier function and the target period, determining seasonal traffic flow affected by seasons corresponding to the prediction time period; The target traffic volume is adjusted based on the seasonal traffic volume.

[0075] Since in actual scenarios, traffic flow may change periodically over time, in an embodiment of the present application, the electronic device may determine seasonal traffic flow affected by seasons during a forecast period.

[0076] Specifically, the electronic device can obtain the Fourier coefficients saved for the scenic spot, determine the Fourier function corresponding to the Fourier coefficients, and determine the cycle to which the predicted time period belongs. In one possible implementation, the electronic device pre-saves the cycle to which each time period of the scenic spot belongs, and the electronic device can determine the target cycle to which the predicted time period belongs. In another possible implementation, the electronic device can also save the time period of the initial cycle of the scenic spot and the time interval between each cycle, and determine the target cycle corresponding to the predicted time period based on the time period and the time interval.

[0077] After determining the target period, the electronic device can determine the seasonal traffic flow affected by the season corresponding to the predicted time period based on the Fourier function and the target period, and adjust the target traffic flow based on the seasonal traffic flow. In a possible implementation, the sum of the target traffic flow and the seasonal traffic flow can be determined, and the target traffic flow is updated using the sum.

[0078] Time series usually show seasonal changes with seasonal changes such as days, weeks, months, and years, also known as periodic changes. For periodic functions, the sine and cosine functions can be immediately associated. In the embodiment of the present application, the seasonal trend is represented by the Fourier series function of sine and cosine: Assuming f(x) is a function with a period of 2π, then its Fourier series is

[0079] ; Use Fourier series to simulate the periodicity of time series. Assume that P represents the period of the time series, P=365.25 represents a yearly period, and P=7 represents a weekly period. Its Fourier series are in the form of: ; In the prediction of traffic flow in scenic spots, this application needs to introduce the concept of lunar month. The year is divided into ordinary years and leap years. Ordinary years have twelve months, and leap years have thirteen months. Months are divided into long months and short months. Long months have thirty days, and short months have twenty-nine days. The average calendar month is equal to a synodic month. The lunar calendar is based on the cycle of the moon's waxing and waning. A synodic month is one month, about 29.53 days, that is, P=29.53. The prediction is made in cooperation with the Gregorian calendar year and week, and three layers of seasonal trends are used for superposition calculation, namely: ; Among them, N, and It is the Fourier coefficient determined in advance based on the historical traffic volume of the scenic spot.

[0080] Figure 4 A schematic diagram of a process for determining seasonal traffic flow provided in an embodiment of the present application, the process comprising the following steps: S401: Acquire Fourier coefficients saved for the scenic spot, and determine the Fourier function corresponding to the Fourier coefficients.

[0081] S402: Determine a target period corresponding to the predicted time period.

[0082] S403: Based on the Fourier function and the target period, determine the seasonal traffic flow affected by the season corresponding to the prediction time period.

[0083] Scenic area traffic flow prediction technology refers to the technology that uses various methods and models to analyze and predict the traffic flow in different time periods, different areas, and different scenes in the scenic area based on historical data, real-time data, external factors, etc. The current development status of scenic area traffic flow prediction technology is that there have been many studies and applications at home and abroad. The main methods used include four-stage method, land use method, traffic characteristic model method, linear prediction method, nonlinear prediction method, wavelet analysis method, neural network method, long short-term memory network method, etc. The defects and possible reasons of scenic area traffic flow prediction technology are that the current methods and models still have the following problems: the prediction accuracy is not high, which cannot meet the needs of refined management and is affected by factors such as data quality, data volume, data source, and data processing; the prediction range and time span are limited, and cannot adapt to changes in the internal and external environment of the scenic area, and are restricted by factors such as the scale, structure, function, and type of the scenic area; the prediction methods and models are relatively single, and cannot fully utilize the advantages of multi-source data and multiple technologies, and are limited by factors such as computing power, algorithm complexity, and model parameters; the prediction results lack effective combination with actual operation management, and cannot achieve dynamic adjustment and optimization, and are interfered by factors such as human interference, emergency events, and emergency response. This application uses historical traffic flow data, related factor information data, etc., through time series analysis and correlation analysis, to construct a traffic flow prediction model, and combines the characteristics of the scenic area and influencing factors to perform dynamic adjustments and optimization to improve prediction accuracy and practicality.

[0084] The embodiment of the present application intends to design a fitting algorithm based on time series decomposition and machine learning, which can handle the situation where there are some outliers in the time series, can also handle the situation where there are some missing values, and can also almost automatically predict the future trend of the time series.

[0085] In the field of time series analysis, there is a common analysis method called time series decomposition. Divided into several parts, namely trend items , Seasonal Item , the remaining items . That is to say, for all t ≥ 0, we have: ; In addition to the addition form, there is also the multiplication form, which is: ; In a possible implementation, when predicting the model, the logarithm is taken first, and then the time series is decomposed to obtain the multiplication form. Based on this method, necessary improvements and optimizations can be made.

[0086] Generally speaking, in real life and production, in addition to seasonal items, trend items, and residual items, there are usually holiday effects. Therefore, when constructing the algorithm, the above four items are considered at the same time, that is: ; in, represents a trend item, which represents the non-periodic trend of the time series, that is, the predicted traffic flow described in the embodiment of the present application; represents a holiday item, that is, the influence weight described in the embodiment of the present application; represents a periodic item, that is, the seasonal traffic flow provided in the embodiment of the present application; It represents the error term or the residual term, that is, the residual vehicle flow provided in the embodiment of the present application. By fitting these terms, the vehicle flow in the predicted time period can be obtained.

[0087] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0088] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0089] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0090] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0091] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for operation management and control of a connected vehicle, characterized in that: It includes: Receive a prediction request, and obtain an identifier of a scenic spot to be predicted and a prediction time period carried in the prediction request; Based on the autoregressive integrated moving average ARIMA model pre-constructed for the identified scenic spot and the traffic volume of each historical time period of the scenic spot, predicting the predicted traffic volume of the scenic spot in the predicted time period; Determine the number of days that include holidays in the forecast time period, obtain random numbers that are the same as the number of days, and determine the impact weight of holidays on the traffic flow in the forecast time period according to the sum of each random number obtained; wherein the greater the number of days, the greater the impact weight; The target traffic flow of the scenic spot in the predicted time period is determined according to the predicted traffic flow and the impact weight.

2. The method according to claim 1, characterized in that The obtaining of random numbers having the same number as the number of days comprises: According to the number of days of the holiday, determine the index value corresponding to the number of days; In a normal distribution whose mean is a preset value and whose standard deviation is the indicator value, determine a number of random numbers that is the same as the number of days.

3. The method according to claim 2, characterized in that The step of determining the index value corresponding to the number of holidays includes: Obtain the maximum and minimum number of holidays in the year to which the forecast time period belongs, and determine the difference between the maximum and minimum number of holidays; According to the ratio of the number of days containing holidays in the predicted time period to the difference, the index value corresponding to the number of days is determined.

4. The method according to any one of claims 1 to 3, characterized in that: After determining the number of days including holidays in the predicted time period and before obtaining the same number of random numbers as the number of days, the method further includes: Obtain the number of days affected by the time interval between the predicted time period and holidays within a preset time interval, obtain the sum of the affected days and the days, and use the sum to adjust the days.

5. The method according to claim 1, characterized in that The method further comprises: Obtain each historical ticket price for each historical time period saved for the scenic spot, and the ticket price for the predicted time period, and obtain each historical weather for each historical time period saved for the scenic spot; Input each historical ticket price, the ticket price, each historical weather and the traffic volume in each historical time period into a pre-trained traffic volume prediction model, and obtain the residual traffic volume affected by the ticket price and weather output by the traffic volume prediction model; The target vehicle flow rate is adjusted based on the residual vehicle flow rate.

6. The method according to claim 5, characterized in that The traffic flow prediction model is trained in the following way: Obtain any data set in the training set, wherein the data set includes ticket prices for multiple historical time periods, ticket prices for predicted time periods, weather for the multiple historical time periods, and vehicle flows for the multiple historical time periods, and obtain the labeled vehicle flows for the predicted time periods affected by weather and ticket prices saved for the data set; Inputting the data group into an original vehicle flow prediction model to obtain an output vehicle flow output by the original vehicle flow prediction model; The original vehicle flow prediction model is trained based on the labeled vehicle flow and the output vehicle flow.

7. The method according to claim 1, characterized in that The method further comprises: Acquire the Fourier coefficients saved for the scenic spot, determine the Fourier function corresponding to the Fourier coefficients; and determine the target period corresponding to the predicted time period; Based on the Fourier function and the target period, determining seasonal traffic flow affected by seasons corresponding to the prediction time period; The target traffic volume is adjusted based on the seasonal traffic volume.

Images