Training method of travel model and vehicle

Abstract

The invention provides a travel model training method and a vehicle, and relates to the field of vehicles. The method comprises the steps of obtaining training data; wherein the training data comprises historical travel data of the vehicle and historical data corresponding to historical variables; the historical travel data comprises historical travel time and historical setting parameters of vehicle functions; the historical variables indicate variables influencing the historical travel data; optimizing the second optimization algorithm based on the first optimization algorithm and the t distribution to obtain a target second optimization algorithm; wherein the target second optimization algorithm is used for determining a target variable according to a historical variable; based on the target second optimization algorithm and the training data, training a to-be-trained travel model to obtain a trained target travel model; wherein the target travel model is used for predicting target travel data of the vehicle. According to the method, the optimization efficiency of the historical variables can be improved, so that the accuracy of predicting the travel data is improved.

Description

Technical Field

[0001] The present application relates to the field of vehicles, and more specifically, to a training method for a travel model and a vehicle in the field of vehicles. Background Art

[0002] Nowadays, cars have become one of the important tools for people's daily travel. With the continuous advancement of technology, intelligent and personalized services have gradually become the key to improving user experience. Based on the user's historical car usage data, predicting the user's travel time and setting the vehicle air conditioning and seat settings in advance are important directions for the development of smart cars.

[0003] Existing prediction methods cannot take into account both global search and local search at the same time during the parameter optimization process. The determined parameters are only local optimal, resulting in low optimization efficiency, making it impossible to accurately predict data such as travel time and service settings. Therefore, how to improve the optimization efficiency of variables to improve the prediction accuracy of the model is an urgent problem to be solved. Summary of the invention

[0004] The present application provides a travel model training method and vehicle, which can improve the efficiency of optimizing historical variables to improve the accuracy of predicting travel data.

[0005] In a first aspect, a method for training a travel model is provided, the method comprising:

[0006] Acquire training data; wherein the training data includes historical travel data of the vehicle and historical data corresponding to historical variables; the historical travel data includes historical travel time and historical setting parameters of vehicle functions; the historical variables indicate variables that have an impact on the historical travel data;

[0007] The second optimization algorithm is optimized based on the first optimization algorithm and the t distribution to obtain a target second optimization algorithm; wherein the first optimization algorithm and the second optimization algorithm are used to represent different heuristic optimization algorithms; the target second optimization algorithm is used to determine the target variable according to the historical variable;

[0008] Based on the target second optimization algorithm and the training data, the travel model to be trained is trained to obtain a trained target travel model; wherein the target travel model is used to predict the target travel data of the vehicle; the target travel data includes the target travel time and the target setting parameters of the vehicle function.

[0009] In the above technical solution, the second optimization algorithm is optimized by using the first optimization algorithm and t distribution to obtain a target second optimization algorithm for determining the target variable according to the historical variable; the travel model to be trained is trained according to the target second optimization algorithm and the historical travel data of the vehicle obtained and the historical data corresponding to the historical variables to obtain a trained target travel model, which is used to predict the target travel data of the vehicle. In this solution, the travel model to be trained is trained by using the optimized target second optimization algorithm and the historical data obtained to obtain a target travel model; since the optimized target second optimization algorithm can improve the accuracy of searching for the target variable, when the target travel model is trained in combination with the target second optimization algorithm, the target travel model can improve the accuracy of the target travel model in predicting the target travel data on the basis of being able to predict the travel data of the vehicle.

[0010] In combination with the first aspect, in some possible implementations, the second optimization algorithm includes a first search strategy, and the second optimization algorithm is optimized based on the first optimization algorithm and the t-distribution to obtain a target second optimization algorithm, including:

[0011] Optimizing the first search strategy by using a first optimization algorithm to obtain an optimized first search strategy;

[0012] Based on the optimized first search strategy and t-distribution, a target second optimization algorithm is obtained.

[0013] In the above technical solution, the first search strategy in the second optimization algorithm is optimized by using the first optimization algorithm, and the target second optimization algorithm is obtained by combining the optimized first search strategy and t distribution. In the prior art, only using the second optimization algorithm to determine the target variable may result in the target variable being determined to be only the optimal variable in the local range, rather than the optimal variable in the global range, and the accuracy of searching the target variable is low, which in turn leads to poor prediction performance of the trained travel model; in order to solve the above problem, this solution optimizes the second optimization algorithm in combination with the first optimization algorithm to enhance the algorithm's ability to search for the optimal variable in a larger range and avoid the search results falling into the local optimum; the second optimization algorithm is optimized in combination with the t distribution, and the disturbance to the variable is dynamically adjusted to achieve a balance between global search and local search, thereby improving the processing performance and efficiency of the target second optimization algorithm in the process of determining the target variable.

[0014] In combination with the first aspect and the above implementation, in some possible implementations, based on the target second optimization algorithm and the training data, the travel model to be trained is trained to obtain the trained target travel model, including:

[0015] The historical variables are iteratively processed through the target second optimization algorithm to obtain the target variable;

[0016] Based on the target variable and historical travel data, the travel model to be trained is trained to obtain the trained target travel model.

[0017] In the above technical solution, the optimized target second optimization algorithm is used to iteratively process the historical variables that have an impact on the travel data to search for the target variable; and the travel model to be trained is trained by the target variable and the historical travel data to obtain the trained target travel model. The target second optimization algorithm can determine a better target variable, that is, a target variable that has a higher degree of influence on the travel data, based on the existing historical variables through continuous iterative search, to avoid the low accuracy of the predicted travel data due to inaccurate selection of the target variable.

[0018] In combination with the first aspect and the above implementations, in some possible implementations, the historical variables are iteratively processed by the target second optimization algorithm to obtain the target variable, including:

[0019] Get the current iteration number i and the variable corresponding to the i-th iteration; where i is a positive integer; when i is 1, the variable corresponding to the i-th iteration is a historical variable;

[0020] Determining a target search strategy in the target second optimization algorithm based on the current iteration number i and the preset number;

[0021] Through the target search strategy and t distribution, the variables corresponding to the i-th iteration are updated to obtain the updated variables corresponding to the i+1-th iteration;

[0022] Based on the updated variable corresponding to the i+1th iteration, the target variable is determined.

[0023] In the above technical solution, by judging the relationship between the current iteration number i and the preset number, the target search strategy that needs to be executed for the variable corresponding to the i-th iteration is determined; and the variable corresponding to the i-th iteration is further updated according to the target search strategy and t distribution to obtain the variable input for the next iteration. Through continuous iterative updates, the variables determined in each iteration can be optimized multiple times to obtain the optimal target variable, thereby improving the efficiency and accuracy of searching the target variable.

[0024] In combination with the first aspect and the above implementation, in some possible implementations, the second optimization algorithm further includes a second search strategy, a third search strategy, and a fourth search strategy;

[0025] Based on the current iteration number i and the preset number, a target search strategy is determined in the target second optimization algorithm, including:

[0026] If the current number of iterations i is less than or equal to the preset number, the second search strategy or the third search strategy is determined as the target search strategy;

[0027] If the current iteration number i is greater than the preset number, the optimized first search strategy or the fourth search strategy is determined as the target search strategy.

[0028] In the above technical scheme, the current number of iterations i is compared with the preset number; when i is less than or equal to the preset number, it indicates that the current iteration is in the early stage, and the second search strategy or the third search strategy is determined as the target search strategy to facilitate determining the variable from a larger range; when i is greater than the preset number, it indicates that the current iteration is in the late stage, and the optimized first search strategy or the fourth search strategy is determined as the target search strategy to facilitate determining the variable from a smaller range; the flexibility of the iterative search in the target second optimization algorithm is improved, and the accuracy of searching the target variable is improved.

[0029] In combination with the first aspect and the above implementations, in some possible implementations, determining the second search strategy or the third search strategy as the target search strategy includes:

[0030] Obtaining a first random number and a first target numerical range; wherein the first random number belongs to the first target numerical range;

[0031] Based on the first target value range, determining a first target threshold value;

[0032] If the first random number is less than the first target threshold, determining the second search strategy as the target search strategy;

[0033] If the first random number is greater than or equal to the first target threshold, the third search strategy is determined as the target search strategy.

[0034] In the above technical solution, if the current number of iterations i is less than or equal to the preset number, a first random number is determined from the first target numerical range. If the first random number is less than the first target threshold determined based on the first target numerical range, the second search strategy is determined as the target search strategy; if the first random number is greater than or equal to the first target threshold, the third search strategy is determined as the target search strategy. During the operation of the algorithm, the deterministic algorithm may encounter specific input data that causes the algorithm to operate inefficiently; by introducing the first random number to determine the target search strategy, the operating efficiency of the target second optimization algorithm can be improved; it can also reduce the probability of deadlock, help the algorithm jump out of the local optimal solution, and explore a broader solution space to determine a more accurate target variable.

[0035] In combination with the first aspect and the above implementations, in some possible implementations, determining the optimized first search strategy or the fourth search strategy as the target search strategy includes:

[0036] Obtaining a second random number; wherein the second random number belongs to a second target numerical range;

[0037] Based on the second target value range, determining a second target threshold value;

[0038] If the second random number is less than the second target threshold, determining the optimized first search strategy as the target search strategy;

[0039] If the second random number is greater than or equal to the second target threshold, the fourth search strategy is determined as the target search strategy.

[0040] In the above technical solution, if the current iteration number i is greater than the preset number, a second random number is determined from the second target numerical range. If the second random number is less than the second target threshold determined based on the second target numerical range, the optimized first search strategy is determined as the target search strategy; if the second random number is greater than or equal to the second target threshold, the fourth search strategy is determined as the target search strategy. The operating efficiency of the target second optimization algorithm is improved; the probability of deadlock can also be reduced, helping the algorithm to jump out of the local optimal solution and explore a wider solution space to determine a more accurate target variable.

[0041] In combination with the first aspect and the above implementation, in some possible implementations, the variable corresponding to the i-th iteration is updated through the target search strategy and the t-distribution to obtain the updated variable corresponding to the i+1-th iteration, including:

[0042] Update the variable corresponding to the i-th iteration through the target search strategy to obtain the first candidate variable;

[0043] Determine a first fitness value of the variable corresponding to the i-th iteration and a second fitness value of the first candidate variable;

[0044] Determining a second candidate variable based on the first fitness value and the second fitness value;

[0045] The second candidate variable is updated through t distribution to obtain the updated variable corresponding to the (i+1)th iteration.

[0046] In the above technical solution, the variable corresponding to the i-th iteration is first updated through the target search strategy to obtain the first candidate variable; and the first fitness value of the variable corresponding to the i-th iteration and the second fitness value of the first candidate variable are determined; the second candidate variable is determined by comparing the first fitness value and the second fitness value; and the second candidate variable is updated again through the t distribution to obtain the updated variable corresponding to the i+1-th iteration. By evaluating the characteristics of the variable through the fitness value, the optimal variable can be selected from the variables determined by the two iterations; the variable is updated in sequence through the target search strategy and the t distribution, and the accuracy of the search variable is further improved.

[0047] In combination with the first aspect and the above implementations, in some possible implementations, determining the target variable based on the updated variable corresponding to the (i+1)th iteration includes:

[0048] If the current iteration number i is greater than the target iteration number, the variable corresponding to the i+1th iteration is determined as the target variable; and / or, if the fitness value of the variable corresponding to the i+1th iteration reaches a preset fitness threshold, the variable corresponding to the i+1th iteration is determined as the target variable.

[0049] In the above technical solution, after the second candidate variable is updated through t distribution and the updated variable corresponding to the i+1th iteration is obtained, the current iteration number i+1 is compared with the target iteration number. If i+1 is greater than the target iteration number, it indicates that the iteration number has reached the preset number requirement, and the variable obtained this time is the final result of the algorithm output, that is, the variable corresponding to the i+1th iteration is determined as the target variable; and / or, the fitness value of the variable corresponding to the i+1th iteration is judged. If the fitness value reaches the preset fitness threshold, it indicates that the variable corresponding to the i+1th iteration is already a variable that meets the requirements, and the variable corresponding to the i+1th iteration can be determined as the target variable. When it is detected that the variable meets the preset requirements, the algorithm operation is terminated and the final variable is determined to ensure that the target variable output by the algorithm is the optimal variable that meets the iteration requirements.

[0050] In a second aspect, a method for predicting travel data is provided, wherein the method comprises:

[0051] Get the target data corresponding to the target variable;

[0052] Inputting the target data into the target travel model to obtain the target travel data of the vehicle; wherein the target travel data includes the target travel time and the target setting parameters for setting the vehicle functions at the target travel time;

[0053] Based on target setting parameters, vehicle functions are controlled at a preset time before the target travel time;

[0054] Wherein, the training method of the target travel model is the method as claimed in any one of claims 1 to 9.

[0055] In the above technical solution, the target data corresponding to the acquired target variable is input into the trained target travel model, and the target travel time of the vehicle and the target setting parameters for the vehicle functions that need to be set at the target travel time are set by the target travel model; and when the time reaches the preset time before the target travel time, the corresponding vehicle functions of the vehicle are controlled according to the target setting parameters. The trained target travel model is used for data prediction to ensure the accuracy of the prediction of travel data; the vehicle functions are controlled at the preset time before the target travel time, which can save time for users' travel. The target setting parameters are predicted according to the target travel model, which can ensure that the setting parameters of the vehicle functions can meet the user's usage habits and improve the user's travel comfort.

[0056] In a third aspect, a travel model training device is provided, the device comprising:

[0057] The first acquisition module is used to acquire training data; wherein the training data includes historical travel data of the vehicle and historical data corresponding to historical variables; the historical travel data includes historical travel time and historical setting parameters of vehicle functions; the historical variables indicate variables that have an impact on the historical travel data;

[0058] An optimization module, used for optimizing the second optimization algorithm based on the first optimization algorithm and the t distribution to obtain a target second optimization algorithm; wherein the target second optimization algorithm is used to determine a target variable according to the historical variable;

[0059] The training module is used to train the travel model to be trained based on the target second optimization algorithm and the training data to obtain the trained target travel model; wherein the target travel model is used to predict the target travel data of the vehicle; the target travel data includes the target travel time and the target setting parameters of the vehicle function.

[0060] In conjunction with the third aspect, in some possible implementations, the second optimization algorithm includes a first search strategy;

[0061] The optimization module is also used to optimize the first search strategy by using a first optimization algorithm to obtain an optimized first search strategy;

[0062] Based on the optimized first search strategy and t-distribution, a target second optimization algorithm is obtained.

[0063] In combination with the third aspect and the above implementations, in some possible implementations, the training module is further used to iteratively process the historical variables through the target second optimization algorithm to obtain the target variables;

[0064] Based on the target variable and historical travel data, the travel model to be trained is trained to obtain the trained target travel model.

[0065] In combination with the third aspect and the above implementation, in some possible implementations, the training module is further used to obtain the current iteration number i and the variable corresponding to the i-th iteration; wherein i is a positive integer; when i is 1, the variable corresponding to the i-th iteration is a historical variable;

[0066] Determining a target search strategy in the target second optimization algorithm based on the current iteration number i and the preset number;

[0067] Through the target search strategy and t distribution, the variables corresponding to the i-th iteration are updated to obtain the updated variables corresponding to the i+1-th iteration;

[0068] Determine the target variable based on the updated variable corresponding to the i+1th iteration.

[0069] In combination with the third aspect and the above implementation, in some possible implementations, the second optimization algorithm further includes a second search strategy, a third search strategy, and a fourth search strategy; the training module is further configured to determine the second search strategy or the third search strategy as the target search strategy if the current number of iterations i is less than or equal to a preset number;

[0070] If the current iteration number i is greater than the preset number, the optimized first search strategy or the fourth search strategy is determined as the target search strategy.

[0071] In combination with the third aspect and the above implementation, in some possible implementations, the training module is further used to obtain a first random number and a first target numerical range; wherein the first random number belongs to the first target numerical range;

[0072] Based on the first target value range, determining a first target threshold value;

[0073] If the first random number is less than the first target threshold, determining the second search strategy as the target search strategy;

[0074] If the first random number is greater than or equal to the first target threshold, the third search strategy is determined as the target search strategy.

[0075] In combination with the third aspect and the above implementation, in some possible implementations, the training module is further used to obtain a second random number and a second target numerical range; wherein the second random number belongs to the second target numerical range;

[0076] Based on the second target value range, determining a second target threshold value;

[0077] If the second random number is less than the second target threshold, determining the optimized first search strategy as the target search strategy;

[0078] If the second random number is greater than or equal to the second target threshold, the fourth search strategy is determined as the target search strategy.

[0079] In combination with the third aspect and the above implementations, in some possible implementations, the training module is further used to update the variable corresponding to the i-th iteration through a target search strategy to obtain a first candidate variable;

[0080] Determine a first fitness value of the variable corresponding to the i-th iteration and a second fitness value of the first candidate variable;

[0081] Determining a second candidate variable based on the first fitness value and the second fitness value;

[0082] The second candidate variable is updated through t distribution to obtain the updated variable corresponding to the (i+1)th iteration.

[0083] In combination with the third aspect and the above-mentioned implementation methods, in some possible implementation methods, the training module is also used to determine the variable corresponding to the i+1th iteration as the target variable if the current number of iterations i is greater than the target number of iterations; and / or, if the fitness value of the variable corresponding to the i+1th iteration reaches a preset fitness threshold, determine the variable corresponding to the i+1th iteration as the target variable.

[0084] In a fourth aspect, a travel data prediction device is provided, the device comprising:

[0085] The second acquisition module is used to acquire target data corresponding to the target variable;

[0086] A prediction module, used for inputting target data into a target travel model to obtain target travel data of the vehicle; wherein the target travel data includes a target travel time and a target setting parameter for setting a vehicle function at the target travel time;

[0087] a control module for controlling vehicle functions at a preset time before a target travel time based on target setting parameters;

[0088] Wherein, the training method of the target travel model is the method as claimed in any one of claims 1 to 9.

[0089] In a fifth aspect, a vehicle is provided, comprising a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, so that the vehicle executes the method in the first aspect or any possible implementation of the first aspect.

[0090] In a sixth aspect, a computer program product is provided, which includes: a computer program code, which, when executed on a computer, enables the computer to execute the method in the first aspect or any possible implementation of the first aspect.

[0091] In the seventh aspect, a computer-readable storage medium is provided, which stores a computer program code. When the computer program code runs on a computer, the computer executes the method in the above-mentioned first aspect or any possible implementation of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0092] Figure 1 is a schematic flow chart of a traditional Skyhawk optimization algorithm provided in an embodiment of the present application;

[0093] Figure 2 is a schematic flow chart of a travel model training method provided in an embodiment of the present application;

[0094] Figure 3 is a schematic flow chart of determining a target variable using a target optimization algorithm provided in an embodiment of the present application;

[0095] Figure 4 is a schematic flow chart of a travel data prediction method provided in an embodiment of the present application;

[0096] Figure 5 is a schematic flow chart of another travel data prediction method provided in an embodiment of the present application;

[0097] Figure 6 It is a structural schematic diagram of a travel model training device provided in an embodiment of the present application;

[0098] Figure 7 It is a structural schematic diagram of a travel data prediction device provided in an embodiment of the present application;

[0099] Figure 8 It is a structural schematic diagram of a vehicle provided in an embodiment of the present application. DETAILED DESCRIPTION

[0100] The technical solution in the present application will be described clearly and in detail below in conjunction with the accompanying drawings. In the description of the embodiments of the present application, unless otherwise specified, " / " means or, for example, A / B can mean A or B: "and / or" in the text is only a description of the association relationship of associated objects, indicating that there can be three relationships, for example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, in the description of the embodiments of the present application, "multiple" means two or more than two.

[0101] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features.

[0102] In modern life, cars have become one of the important tools for people's daily travel. With the continuous advancement of science and technology, intelligent and personalized services have gradually become the key to improving user experience. Based on the user's historical car usage data, predicting the user's travel time and setting the vehicle air conditioning and seat settings in advance is an important direction for the development of smart cars. This data-driven personalized service can not only improve the user's driving experience, but also promote energy conservation and emission reduction and enhance safety. The development of smart car technology enables vehicles to automatically adjust the internal environment and improve comfort according to the needs and preferences of users. At the same time, users' demand for cars is no longer limited to basic travel functions, but more focused on driving experience and personalized services. Therefore, analyzing users' car usage habits, predicting their travel patterns, and then providing customized services is an effective way to meet users' diverse needs.

[0103] The research and implementation of this project has important social and economic significance. First, by predicting the user's travel time and setting the vehicle environment in advance, the user's driving experience can be greatly improved, making travel more comfortable and convenient. Secondly, by intelligently adjusting equipment such as air conditioning and seats, it can reach the optimal state before the user uses it, reducing energy waste, and conforming to the concept of green and environmentally friendly travel. In addition, by predicting the user's travel pattern, the vehicle can be inspected and maintained in advance to reduce the risk of accidents. At the same time, the project can deepen the analysis and application of user data, provide valuable data support for automakers and service providers, and help them better understand market demand and optimize product design and service strategies. Ultimately, this will promote the development of smart car technology, bring new growth points to the automotive industry, and promote the upgrading of the entire industrial chain.

[0104] In the process of predicting travel data, the model is usually trained based on the Long Short-Term Memory Network (LSTM), combined with the vehicle-connected data to predict vehicle travel; in the training process of the LSTM model, finding the optimal parameters is a key step. In the existing technology, the traditional Aquila Optimizer (AO) is usually used to find the optimal parameters.

[0105] Figure 1 is a schematic flow chart of a traditional Skyhawk optimization algorithm provided in an embodiment of the present application. Figure 1As shown, the Sky Eagle optimization algorithm includes a first search strategy, a second search strategy, a third search strategy and a fourth search strategy; the strategy for updating the Sky Eagle position is determined by the current number of iterations i and the target number of iterations K; when i≤K, iterative update is performed; when i<2 / 3K and the random number rand<0.5, it is determined to update the Sky Eagle position through the second search strategy; when i<2 / 3K and the random number rand≥0.5, it is determined to update the Sky Eagle position through the third search strategy; when i≥2 / 3K and the random number rand<0.5, it is determined to update the Sky Eagle position through the first search strategy; when i≥2 / 3K and the random number rand≥0.5, it is determined to update the Sky Eagle position through the fourth search strategy; if the current number of iterations i reaches the target number of iterations K, the iterative process is terminated.

[0106] Specifically, the first search strategy corresponds to the expansion and development stage, in which the Skyhawk determines the prey area and can perform low-altitude flight and vertical attack to update the position; the corresponding mathematical model is as shown in Formula 1:

[0107]

[0108] Among them, α and β represent development adjustment parameters, taking smaller values ​​(0,1), UB is the upper bound, and LB is the lower bound; represents the individual position of the Tianying optimization algorithm in the t+1th iteration; represents the best individual position obtained by the algorithm up to the tth iteration; represents the average position of the population at the tth iteration, rand is a random number in [0, 1].

[0109] The second search phase corresponds to the expansion exploration phase, in which the eagle will determine the prey area and update its position by vertically lying prone and flying high in the sky. The corresponding mathematical model is as shown in Formula 2:

[0110]

[0111] Where T represents the maximum number of iterations of the algorithm.

[0112] The third search phase corresponds to the reduced exploration phase, in which the Skyhawk uses short gliding combined with contour flight to update the position. The corresponding mathematical model is as shown in Formula 3:

[0113]

[0114] Where, Levy(D) is the flight distribution function; A random number in the range [1, N].

[0115] The fourth search phase corresponds to the narrowing development phase, in which the eagle uses the escape trajectory of the prey to attack it to update the eagle's position. The corresponding mathematical model is shown in Formula 4:

[0116]

[0117] Where QF(t) represents the quality function value used to balance the search strategy, G represents the various movements of AO in the process of tracking prey. G1 and G2 represent the linearly decreasing flight slope values, ranging from [0,2]. G1 = 2 × rand-1.

[0118] Although the traditional Sky Eagle optimization algorithm has the advantages of strong optimization ability and simple process, the AO algorithm often has limitations when searching for the optimal solution. For example, it cannot take into account both global search ability and local search ability at the same time, which may cause the algorithm to fall into local optimality when dealing with practical problems, affecting the accuracy of the prediction results.

[0119] Therefore, an embodiment of the present application provides a method for training a travel model, which takes the traditional Sky Eagle optimization algorithm as the second optimization algorithm to be optimized and optimizes it to improve the efficiency of optimizing historical variables and enhance the accuracy of predicting travel data.

[0120] Figure 2 It is a schematic flow chart of a method for training a travel model provided in an embodiment of the present application. It should be understood that the method can be applied to vehicles.

[0121] It should be noted that in order to avoid excessive local computing in the vehicle, this method can also be applied to servers and cloud platforms associated with the vehicle.

[0122] For example, Figure 2 As shown, the method 200 includes:

[0123] S201, obtaining training data; wherein the training data includes historical travel data of the vehicle and historical data corresponding to historical variables; the historical travel data includes historical travel time and historical setting parameters of vehicle functions; and the historical variables indicate variables that have an impact on the historical travel data.

[0124] Vehicle network data records various key parameters of the vehicle during operation, such as speed, acceleration, fuel consumption, engine status, vehicle operating time, and the parameters set for vehicle functions when the user rides the vehicle. Vehicle network data can not only reflect the status of the vehicle, but also contain rich historical information. Based on a variety of historical information, it is possible to analyze and predict vehicle behavior. In the process of training the travel model, multiple historical travel times of the vehicle are obtained, historical setting parameters for setting vehicle functions at historical travel times, historical variables that affect historical travel data, and corresponding historical data.

[0125] Exemplarily, vehicle functions may include air conditioning, seats, music playback, etc.; historical variables may include weather, festivals, holidays, and seasons, etc.

[0126] It should be noted that in order to improve the accuracy of travel data prediction, historical travel time can be accurate to a specific time or time period on a certain day.

[0127] In one implementation, the multiple sets of training data of the vehicle obtained may be presented in the form of a correspondence table.

[0128] Exemplarily, the corresponding relationship between the historical travel time, the historical setting parameters and the historical data corresponding to the historical variables is shown in Table 1.

[0129] Table 1

[0130]

[0131] As shown in Table 1, the acquired training data includes that the vehicle traveled at 9:00 on July 2, the air-conditioning temperature was adjusted to 21°C, the seat position was adjusted to position A, and music was played; that day was a weekday, hot weather, and summer; the vehicle traveled at 8:50 on July 3, the air-conditioning temperature was adjusted to 23°C, and music was played; that day was a weekday, sunny weather; the vehicle traveled at 13:00 on July 4, music was played, and the windows were opened 20%; that day was a holiday, and cloudy weather.

[0132] S202, optimizing the second optimization algorithm based on the first optimization algorithm and t-distribution to obtain a target second optimization algorithm; wherein the first optimization algorithm and the second optimization algorithm are used to represent different heuristic optimization algorithms; the target second optimization algorithm is used to determine the target variable according to the historical variable.

[0133] Heuristic optimization algorithm is a method for solving optimization problems. It draws on phenomena in nature or human experience to design algorithms to find approximate solutions to problems. These algorithms are usually suitable for solving complex problems that are difficult to find exact solutions using traditional mathematical methods, especially when the solution space is very large or the solution structure is complex. In the process of training the travel model, it is necessary to determine the target variables that have a greater impact on the travel data; in order to determine a more accurate target variable, the heuristic optimization algorithm can be combined to determine the target variable based on the historical variables obtained. Heuristic optimization algorithms can include the Eagle Optimization Algorithm, the Whale Optimization Algorithm, the Simulated Annealing Algorithm, the Genetic Algorithm, and the Particle Swarm Optimization Algorithm.

[0134] It should be noted that the embodiment of the present application takes the first optimization algorithm as the whale optimization algorithm and the second optimization algorithm as the eagle optimization algorithm as an example; in actual application, the first optimization algorithm and the second optimization algorithm can both adopt other heuristic optimization algorithms with the same functions as the whale optimization algorithm and the eagle optimization algorithm; the embodiment of the present application does not limit this.

[0135] Whale Optimization Algorithm (WOA) is a heuristic optimization algorithm designed based on the predation behavior of whales. WOA is a relatively new meta-heuristic optimization algorithm proposed in recent years, which is mainly used to solve continuous optimization problems. The Whale Optimization Algorithm is mainly based on the social behavior and predation strategy of whale groups. Specifically, the algorithm imitates a special predation method of humpback whales - "bubble net attack", which includes several stages: surrounding prey, bubble attack and prey predation.

[0136] In the stage of circling the prey, assuming that the optimal solution is the location of the prey, the behavior of the whale moving towards the location can be seen in Formula 5 and Formula 6:

[0137]

[0138]

[0139] Where t represents the number of iterations; represents the updated coefficient vector, is a random vector in [0, 1]; It decreases linearly from 2 to 0 as the number of iterations increases; represents the best position of the whale so far; Represents the current whale position vector.

[0140] During the bubble attack phase, the whale's bubble net behavior is converted into a digital model through two mechanisms: contraction surround and spiral position update.

[0141] The formula for the shrink-wrap mechanism is shown in Formula 6, which will not be elaborated in this article;

[0142] The spiral position update mechanism is as shown in Formula 7:

[0143]

[0144] in, represents the distance between the whale and its prey; b is a constant value used to identify the shape of a logarithmic spiral, and k is a random number in the range [-1, 1].

[0145] During the prey hunting phase, use The random value of forces the whale to expand its search range, and the whale's position is updated through random positions. The corresponding mathematical models are as follows:

[0146]

[0147]

[0148] in, Represents the position vector of a randomly determined whale in the current population.

[0149] The WOA algorithm starts with a series of random solutions. In each iteration, the whale updates its position according to the randomly selected solution or the current optimal solution. At the same time, WOA freely switches between the shrinking and wrapping mechanism and the spiral position update mechanism according to the random number p in (0,1). When p<0.5, the whale position is updated by the above formula 6; when p≥0.5, the whale position is updated by the above formula 7.

[0150] The second optimization algorithm will replace the previous optimal variable with the current best variable in each iteration, which may cause the variable to miss the global optimum when it falls into the local optimum. In order to solve the above problem, the embodiment of the present application introduces t distribution to perturb the position of the variable.

[0151] The t-distribution (Student's t-distribution) is a continuous probability distribution that is widely used in statistics, especially for estimating the population mean when the sample size is small and the population standard deviation is unknown. The t-distribution was proposed by William Sealy Gosset in the early 20th century, who used the pseudonym "Student" at the time; therefore, the t-distribution is also called the Student distribution. The t-distribution is symmetrical and has a shape similar to the normal distribution, but with thicker tails. The t-distribution is controlled by a parameter, called the degrees of freedom (df). The degrees of freedom determine the shape of the distribution, and as the degrees of freedom increase, the t-distribution gradually tends to the normal distribution; compared to the normal distribution, the t-distribution has a thicker tail, which means that it is more resistant to the influence of outliers and is more suitable for processing small sample data. The peak of the t-distribution is slightly higher than that of the normal distribution, which reflects its estimation of the central tendency of small sample data.

[0152] In order to improve the application effect of t distribution in the target second optimization algorithm, the embodiment of the present application can adopt adaptive t distribution and use the number of iterations as the degree of freedom parameter of t distribution, so as to realize global search and local exploration in the iteration process and accelerate the convergence speed of the algorithm. The implementation method of adaptive t distribution is as shown in Formula 10:

[0153]

[0154] in, is the position of the optimal solution in the jth dimension after the adaptive t-distribution mutation disturbance, is the position of the optimal solution in the jth dimension before the mutation disturbance, and the number of iterations t is used as the degree of freedom parameter of the t distribution.

[0155] Updating the variables again through t-distribution can effectively deal with nonlinear and multimodal optimization problems because it can automatically adjust the intensity of the disturbance during the search process, thereby better balancing the needs of global and local searches.

[0156] In one implementation, the second optimization algorithm is the Eagle optimization algorithm, which includes a first search strategy, which instructs the Eagle to move around the prey in a gradually smaller circular path; the first search strategy is optimized by the whale optimization algorithm to obtain an optimized first search strategy, so that the Eagle moves along a spiral path on the basis of moving in the circular path, avoiding the updated position from falling into a local optimum; and then based on the optimized first search strategy and t-distribution, the target second optimization algorithm is obtained.

[0157] In the embodiment of the present application, the whale optimization algorithm is used to optimize the first search strategy in the second optimization algorithm, and the optimized first search strategy and t distribution are combined to obtain the target second optimization algorithm. In the prior art, only using the second optimization algorithm to determine the target variable may result in the determined target variable being only the optimal variable in the local range, rather than the optimal variable in the global range, and the accuracy of searching the target variable is low, which in turn leads to poor prediction performance of the trained travel model; in order to solve the above problems, this solution optimizes the second optimization algorithm in combination with the whale optimization algorithm to enhance the algorithm's ability to search for the optimal variable in a larger range and avoid the search results falling into the local optimum; the second optimization algorithm is optimized in combination with the t distribution, and the disturbance to the variable is dynamically adjusted to achieve a balance between global search and local search, thereby improving the processing performance and processing efficiency of the target second optimization algorithm in the process of determining the target variable.

[0158] S203, based on the target second optimization algorithm and the training data, the travel model to be trained is trained to obtain a trained target travel model; wherein the target travel model is used to predict the target travel data of the vehicle; the target travel data includes the target travel time and the target setting parameters of the vehicle function.

[0159] Exemplarily, the travel model to be trained can be an LSTM model. The step of finding the optimal parameters in the LSTM model is optimized according to the optimized target second optimization algorithm, and the acquired historical travel data and historical data corresponding to the historical variables are input into the LSTM model, and the model is trained to obtain the trained target travel model; so that the target travel model has the ability to predict the target travel time of the vehicle and the target setting parameters of the vehicle functions.

[0160] In one implementation, historical variables are iteratively processed through a target second optimization algorithm to obtain a target variable; based on the target variable and historical travel data, a travel model to be trained is trained to obtain a trained target travel model.

[0161] The optimized target second optimization algorithm can determine the target variable that has a greater impact on the travel data based on the existing historical variables through continuous updating and iterative processing; after the target variable is determined, the variable data corresponding to the target variable and the historical travel data are input into the LSTM model for training. The model can analyze the corresponding relationship between the target variable and the travel data, so that the trained target travel model can predict the target travel data based on the target data corresponding to the input target variable.

[0162] Exemplarily, the historical variables obtained include weather, date attributes, and seasons; the data corresponding to the weather include high temperature, sunny, cloudy, rainy, etc.; the data corresponding to the date attributes may include holidays, weekdays, weekends, etc.; the data corresponding to the seasons may include spring, summer, autumn, and winter, etc.; the above historical variables are iteratively processed through the target second optimization algorithm, and the variable with the greatest impact on the travel data is the date attribute. Then, the LSTM model can be trained based on the data corresponding to the date attribute and the corresponding historical travel data to obtain the trained target travel data.

[0163] In the embodiment of the present application, the optimized target second optimization algorithm is used to iteratively process the historical variables that have an impact on the travel data to search for the target variable; and the travel model to be trained is trained by the target variable and the historical travel data to obtain the trained target travel model. The target second optimization algorithm can determine a better target variable, that is, a target variable that has a higher degree of impact on the travel data, based on the existing historical variables through continuous iterative search, to avoid the low accuracy of the predicted travel data due to inaccurate selection of the target variable.

[0164] In one implementation, historical variables are iteratively processed through a target second optimization algorithm to obtain target variables, including: obtaining the current iteration number i and the variable corresponding to the i-th iteration; wherein i is a positive integer; when i is 1, the variable corresponding to the i-th iteration is the historical variable; based on the current iteration number i and the preset number, a target search strategy is determined in the target second optimization algorithm; through the target search strategy and t-distribution, the variable corresponding to the i-th iteration is updated to obtain the updated variable corresponding to the i+1-th iteration; based on the updated variable corresponding to the i+1-th iteration, the target variable is determined.

[0165] The preset number of times may be a preset fixed value; or a preset ratio may be set, and the preset number of times may be determined according to the preset ratio and the target number of iterations.

[0166] Exemplarily, the preset number of times may be set to 10 times; or, the target number of iterations may be set to 20 times, the preset ratio may be set to 40%, and the preset number of times may be determined to be 8 times.

[0167] The second optimization algorithm includes a first search strategy, a second search strategy, a third search strategy and a fourth search strategy; the first search strategy in the second optimization algorithm is optimized, and the target second optimization algorithm obtained by combining the t-distribution includes the optimized first search strategy, the second search strategy, the third search strategy, the fourth search strategy and the t-distribution.

[0168] Exemplarily, in the process of iteratively processing historical variables through the target second optimization algorithm, various parameters involved in the algorithm will be initialized, including the setting of the target number of iterations; based on the relationship between the current number of iterations and the preset number of iterations, the target search strategy required for the current number of iterations can be determined from the four search strategies in the target second optimization algorithm; then the variables are updated in turn according to the target search strategy and t-distribution to determine the variables corresponding to the updated i+1th iteration; and based on the variables corresponding to the updated i+1th iteration, the final target variable is determined.

[0169] In the embodiment of the present application, by judging the relationship between the current iteration number i and the preset number, the target search strategy that needs to be executed for the variable corresponding to the i-th iteration is determined; and the variable corresponding to the i-th iteration is further updated according to the target search strategy and t distribution to obtain the variable input for the next iteration. Through continuous iterative updates, the variables determined in each iteration can be optimized multiple times to obtain the optimal target variable, thereby improving the efficiency and accuracy of searching the target variable.

[0170] In one implementation, based on the current number of iterations i and the preset number, the target search strategy is determined in the target second optimization algorithm, including: if the current number of iterations i is less than or equal to the preset number, the second search strategy, or the third search strategy is determined as the target search strategy; if the current number of iterations i is greater than the preset number, the optimized first search strategy, or the fourth search strategy is determined as the target search strategy.

[0171] Exemplarily, the target number of iterations can be set to 9 times, and the preset ratio is set to 2 / 3, then the preset number of times can be 6 times; if the current number of iterations i is less than 6 times, it indicates that it is currently in the early stage of iteration, and the early stage of iteration needs to enter the exploration stage, and the exploration stage includes the second search strategy and the third search strategy, then the second search strategy or the third search strategy can be determined as the target search strategy; if the current number of iterations i reaches 6 times, it indicates that it is currently in the late stage of iteration, and the late stage of iteration needs to enter the development stage, and the development stage includes the optimized first search strategy and the fourth search strategy, then the optimized first search strategy and the fourth search strategy can be determined as the target search strategy.

[0172] In an embodiment of the present application, the current number of iterations i is compared with a preset number; when i is less than or equal to the preset number, it indicates that the current iteration is in the early stage, and the second search strategy or the third search strategy is determined as the target search strategy to facilitate determining variables from a larger range; when i is greater than the preset number, it indicates that the current iteration is in the late stage, and the optimized first search strategy or the fourth search strategy is determined as the target search strategy to facilitate determining variables from a smaller range; the flexibility of the iterative search in the target second optimization algorithm is improved, and the accuracy of searching the target variable is improved.

[0173] In one implementation, the second search strategy or the third search strategy is determined as the target search strategy, including: obtaining a first random number and a first target numerical range; wherein the first random number belongs to the first target numerical range; based on the first target numerical range, determining a first target threshold; if the first random number is less than the first target threshold, determining the second search strategy as the target search strategy; if the first random number is greater than or equal to the first target threshold, determining the third search strategy as the target search strategy.

[0174] The second search strategy and the third search strategy are both used to explore the optimal solution; when determining the specific target search strategy, random numbers can be introduced to avoid multiple updates through the same strategy, which causes the algorithm to fall into a local optimum; a first random number is obtained from the first target numerical range, and a first target threshold corresponding to the first random number is determined based on the first target numerical range, and the first random number is divided into two ranges, so that the probability distribution of adopting the two search strategies is more even.

[0175] Exemplarily, the first target numerical range can be set to [0,1], and the first target threshold can be determined as 0.5; if the first random number obtained is 0.2, the second search strategy is determined as the target search strategy; if the first random number obtained is 0.7, the third search strategy is determined as the target search strategy.

[0176] In the embodiment of the present application, if the current number of iterations i is less than or equal to the preset number, a first random number is determined from the first target numerical range, and if the first random number is less than the first target threshold determined based on the first target numerical range, the second search strategy is determined as the target search strategy; if the first random number is greater than or equal to the first target threshold, the third search strategy is determined as the target search strategy. During the operation of the algorithm, a deterministic algorithm may encounter specific input data that causes the algorithm to operate inefficiently; by introducing the first random number to determine the target search strategy, the operating efficiency of the target second optimization algorithm can be improved; it can also reduce the probability of deadlock, help the algorithm jump out of the local optimal solution, and explore a broader solution space to determine a more accurate target variable.

[0177] In one implementation, the optimized first search strategy, or the fourth search strategy, is determined as the target search strategy, including: obtaining a second random number and a second target numerical range; wherein the second random number belongs to the second target numerical range; based on the second target numerical range, determining a second target threshold; if the second random number is less than the second target threshold, determining the optimized first search strategy as the target search strategy; if the second random number is greater than or equal to the second target threshold, determining the fourth search strategy as the target search strategy.

[0178] The optimized first search strategy and the fourth search strategy are both used to develop the optimal solution; when determining the specific target search strategy, random numbers can also be introduced to avoid multiple updates through the same strategy, causing the algorithm to fall into a local optimum; a second random number is obtained from the second target numerical range, and the second target threshold corresponding to the second random number is determined according to the second target numerical range, and the second random number is divided into two ranges, so that the probability distribution of adopting the two search strategies is more even.

[0179] Exemplarily, the second target numerical range can be set to [0,2], and the second target threshold can be determined as 1; if the second random number obtained is 0.8, the optimized first search strategy is determined as the target search strategy; if the second random number obtained is 1.3, the fourth search strategy is determined as the target search strategy.

[0180] In the embodiment of the present application, if the current iteration number i is greater than the preset number, a second random number is determined from the second target numerical range, and if the second random number is less than the second target threshold value determined based on the second target numerical range, the optimized first search strategy is determined as the target search strategy; if the second random number is greater than or equal to the second target threshold value, the fourth search strategy is determined as the target search strategy. The operating efficiency of the target second optimization algorithm is improved; the probability of deadlock can also be reduced, helping the algorithm to jump out of the local optimal solution and explore a wider solution space to determine a more accurate target variable.

[0181] In one implementation, a variable corresponding to the i-th iteration is updated through a target search strategy and a t-distribution to obtain an updated variable corresponding to the i+1-th iteration, including: updating the variable corresponding to the i-th iteration through a target search strategy to obtain a first candidate variable; determining a first fitness value of the variable corresponding to the i-th iteration and a second fitness value of the first candidate variable; determining a second candidate variable based on the first fitness value and the second fitness value; and updating the second candidate variable through a t-distribution to obtain an updated variable corresponding to the i+1-th iteration.

[0182] Exemplarily, in the process of updating the variable corresponding to the i-th iteration, it is first updated through the target search strategy to obtain the first candidate variable, and the first fitness value of the variable corresponding to the i-th iteration is compared with the second fitness value of the first candidate variable to determine a better variable, and the variable with a higher fitness value is determined as the updated second candidate variable; then the second candidate variable is updated again through the t distribution to determine the updated variable corresponding to the i+1-th iteration; in the next iteration, the variable corresponding to the i+1-th iteration is updated again to obtain the variable corresponding to the i+2-th iteration, and so on, until the end of the algorithm.

[0183] In the process of updating the second candidate variables through t distribution, for each second candidate variable, a disturbance vector is extracted from the t distribution and then added to the second candidate variable to generate a mutation vector; if the fitness of the mutation vector is higher than that of the second candidate variable, the mutation vector is accepted and the degrees of freedom parameter is updated according to its performance; otherwise, the second candidate variable is retained and the degrees of freedom parameter is adjusted accordingly.

[0184] In the embodiment of the present application, the variable corresponding to the i-th iteration is first updated through the target search strategy to obtain the first candidate variable; and the first fitness value of the variable corresponding to the i-th iteration and the second fitness value of the first candidate variable are determined; the second candidate variable is determined by comparing the first fitness value and the second fitness value; and the second candidate variable is updated again through the t distribution to obtain the updated variable corresponding to the i+1-th iteration. By evaluating the characteristics of the variable through the fitness value, the optimal variable can be selected from the variables determined by the two iterations; the variable is updated in sequence through the target search strategy and the t distribution, and the accuracy of the search variable is further improved.

[0185] In one implementation, the target variable is determined based on the updated variable corresponding to the i+1th iteration through t-distribution, including: if the current iteration number i is greater than the target iteration number, the variable corresponding to the i+1th iteration is determined as the target variable; and / or, if the fitness value of the variable corresponding to the i+1th iteration reaches a preset fitness threshold, the variable corresponding to the i+1th iteration is determined as the target variable.

[0186] When the current number of iterations reaches the target number of iterations, it indicates that the current iteration has met the user's needs, and the variable corresponding to the i+1th iteration can be determined as the final target variable; or, a preset fitness threshold can be set for the algorithm result, and the fitness value of the variable corresponding to the i+1th iteration is judged. If the preset fitness threshold is reached, the variable corresponding to the i+1th iteration is determined as the target variable.

[0187] In the embodiment of the present application, after the second candidate variable is updated through t distribution and the updated variable corresponding to the i+1th iteration is obtained, the current iteration number i+1 is compared with the target iteration number. If i+1 is greater than the target iteration number, it indicates that the number of iterations has reached the preset number requirement, and the variable obtained this time is the final result of the algorithm output, that is, the variable corresponding to the i+1th iteration is determined as the target variable; and / or, the fitness value of the variable corresponding to the i+1th iteration is judged. If the fitness value reaches the preset fitness threshold, it indicates that the variable corresponding to the i+1th iteration is already a variable that meets the requirements, and the variable corresponding to the i+1th iteration can be determined as the target variable. When it is detected that the variable meets the preset requirements, the algorithm operation is terminated and the final variable is determined to ensure that the target variable output by the algorithm is the optimal variable that meets the iteration requirements.

[0188] In summary, in the embodiment of the present application, the first optimization algorithm and t distribution are used to optimize the second optimization algorithm to obtain a target second optimization algorithm for determining the target variable according to the historical variable; the travel model to be trained is trained according to the target second optimization algorithm and the historical travel data of the vehicle obtained and the historical data corresponding to the historical variables to obtain a trained target travel model, which is used to predict the target travel data of the vehicle. In this solution, the travel model to be trained is trained by the optimized target second optimization algorithm and the historical data obtained to obtain a target travel model; because the optimized target second optimization algorithm can improve the accuracy of searching for the target variable, when the target travel model is trained in combination with the target second optimization algorithm, the target travel model can improve the accuracy of the target travel model in predicting the target travel data on the basis of being able to predict the travel data of the vehicle.

[0189] Figure 3 It is a schematic flow chart of determining the target variable using the target second optimization algorithm provided in an embodiment of the present application.

[0190] For example, Figure 3 As shown, the process includes:

[0191] S301, setting algorithm parameters of the target second optimization algorithm and initializing variables.

[0192] Among them, the target second optimization algorithm is used to continuously iterate the initial variables to obtain the optimal target variables; in this process, it involves the setting of various judgment conditions, such as the target number of iterations, the target threshold of random numbers, etc.

[0193] In order to avoid program errors or result uncertainties caused by random variables, before running the algorithm, it is necessary to set the algorithm parameters involved and initialize the variables to be processed.

[0194] Exemplarily, the target number of iterations K is set to 30 times, the value range of the random number is set to 0 to 1, and the target threshold of the random number is set to 0.5; an initial solution is set for the target second optimization algorithm, and the initial solution can be a randomly generated variable, or a predetermined variable; for example, weather, festivals, holidays, seasons, etc.

[0195] S302, determine whether the current number of iterations i is less than or equal to the target number of iterations K; if so, execute S303; if not, end the algorithm process.

[0196] Exemplarily, in the target second optimization algorithm, the number of iterations will continue to accumulate; before performing the algorithm operation, it is determined whether the current number of iterations has reached the target number of iterations; if the target number of iterations has been reached, indicating that the iteration requirements have been met, the algorithm is determined to be finished and the iteration process is exited; if the target number of iterations has not been reached, it is necessary to enter the subsequent search phase to update the variables.

[0197] Optionally, the implementation of S302 can refer to Figure 2 The relevant description of S203 in the present application embodiment will not be repeated here.

[0198] S303, determine whether the current iteration number i is less than the preset number; if so, execute S304; if not, execute S307.

[0199] Among them, the preset number of times can be a pre-set fixed value; a preset ratio can also be set, and the preset number of times is determined according to the preset ratio and the target number of iterations; the preset ratio can be set according to actual needs. The embodiment of the present application takes the preset ratio of 2 / 3 as an example.

[0200] For example, if the preset ratio is 2 / 3 and the target number of iterations is 30, then the preset number can be determined to be 20; if the current number of iterations is less than or equal to 20, it indicates that it is currently in the early stage of iteration and the update of variables should enter the exploration stage; if the current number of iterations is greater than 20, it indicates that it is in the late stage of iteration and the update of variables should enter the development stage.

[0201] Optionally, the implementation of S303 can refer to Figure 2 The relevant description of S203 in the present application embodiment will not be repeated here.

[0202] S304, obtaining a first random number rand1, and determining whether the first random number rand1 is less than a first target threshold of the first random number; if so, executing S305; if not, executing S306.

[0203] Exemplarily, when the current number of iterations is less than or equal to a preset number, the search strategy to be adopted for the current variable is determined by judging the relationship between the first random number and the first target threshold, and the search strategy in the exploration phase includes a second search strategy and a third search strategy.

[0204] Optionally, the implementation of S304 can refer to Figure 2 The relevant description of S203 in the present application embodiment will not be repeated here.

[0205] S305, updating the variable by using the second search strategy to obtain an updated variable.

[0206] Exemplarily, the first target numerical range of the first random number is pre-set to [0,1], and the first target threshold of the first random number is 0.5; if the acquired first random number is 0.3, which is less than the first target threshold 0.5, the second search strategy is determined as the strategy for updating the variable, and the variable is updated to obtain the updated variable.

[0207] Optionally, the implementation of S305 can refer to Figure 2 The relevant description of S203 in the present application embodiment will not be repeated here.

[0208] S306, updating the variables through the third search strategy to obtain updated variables.

[0209] Exemplarily, the first target numerical range of the first random number is pre-set to [0,1], and the first target threshold of the first random number is 0.5; if the obtained first random number is 0.9, which is greater than the first target threshold 0.5, the third search strategy is determined as the strategy for updating the variable, and the variable is updated to obtain the updated variable.

[0210] Optionally, the implementation of S306 can refer to Figure 2 The relevant description of S203 in the present application embodiment will not be repeated here.

[0211] S307, obtaining a second random number rand2, and determining whether the second random number rand2 is less than a second target threshold value of the second random number; if so, executing S308; if not, executing S309.

[0212] Exemplarily, when the current number of iterations is greater than a preset number, the search strategy to be adopted for the current variable is determined by judging the relationship between the second random number and the second target threshold, and the search strategy in the development stage includes the optimized first search strategy and the fourth search strategy.

[0213] Optionally, the implementation of S307 can refer to Figure 2The relevant description of S203 in the present application embodiment will not be repeated here.

[0214] S308, updating the variables using the optimized first search strategy to obtain updated variables.

[0215] Exemplarily, the second target numerical range of the second random number is pre-set to [0,1], and the second target threshold of the second random number is 0.5; if the obtained second random number is 0.4, which is less than the second target threshold of 0.5, the optimized first search strategy is determined as the strategy for updating the variable, and the variable is updated to obtain the updated variable.

[0216] Optionally, the implementation of S308 can refer to Figure 2 The relevant description of S203 in the present application embodiment will not be repeated here.

[0217] S309, updating the variables using the fourth search strategy to obtain updated variables.

[0218] Exemplarily, the second target numerical range of the second random number is pre-set to [0,1], and the second target threshold of the second random number is 0.5; if the obtained second random number is 0.6, which is greater than the second target threshold 0.5, the fourth search strategy is determined as the strategy for updating the variable, and the variable is updated to obtain the updated variable.

[0219] Optionally, the implementation of S309 can refer to Figure 2 The relevant description of S203 in the present application embodiment will not be repeated here.

[0220] S310, updating the updated variable through t distribution to obtain the target variable.

[0221] Exemplarily, after the input variables are updated through the search strategy to obtain updated variables, the updated variables are updated again through t distribution to obtain the final target variable.

[0222] Optionally, the implementation of S310 can refer to Figure 2 The relevant description of S203 in the present application embodiment will not be repeated here.

[0223] S311, determine whether to end the algorithm process according to the target variable; if so, end the algorithm process; if not, execute S303.

[0224] Exemplarily, after the target variable is obtained by updating the variable through t-distribution, it can be determined whether the target variable meets the user's requirements, or whether the number of iterations reaches the preset target number of iterations, so as to determine whether to end the iterative update process; if the target variable meets the user's requirements, and / or the current number of iterations reaches the target number of iterations, it is determined to end the algorithm process; if not, return to S303, the current number of iterations is accumulated plus 1, and the variable is updated again until the requirements are met.

[0225] Optionally, the implementation of S311 can refer to Figure 2 The relevant description of S203 in the present application embodiment will not be repeated here.

[0226] In an embodiment of the present application, the initial external variables are continuously searched and optimized through the optimized target second optimization algorithm to obtain target variables that meet preset conditions; during the optimization process, the corresponding search strategy is determined by the current number of iterations and the target number of iterations, and random numbers are introduced to improve the operating efficiency of the target second optimization algorithm; the probability of deadlock is reduced, helping the algorithm to jump out of the local optimal solution and explore a broader solution space to determine a more accurate target variable.

[0227] Figure 4 is a schematic flow chart of a travel data prediction method provided in an embodiment of the present application. It should be understood that the method can be applied to vehicles.

[0228] For example, Figure 4 As shown, the method 400 includes:

[0229] S401, obtaining target data corresponding to the target variable.

[0230] Exemplarily, if the target variable is weather, weather data within a preset time period in the future can be obtained. For example, if the current date is July 20, the weather on July 21 is sunny, the weather on July 22 is heavy rain, and the weather on July 23 is cloudy.

[0231] S402, inputting target data into a target travel model to obtain target travel data of the vehicle; wherein the target travel data includes a target travel time and target setting parameters for setting vehicle functions at the target travel time.

[0232] Based on the target variables and historical travel data, the target travel model is trained so that the target travel model has the ability to predict travel data according to the target variables; in the process of model application, the target data corresponding to the target variables of the future time period are input into the target travel model, and the target travel model can output the corresponding target travel time and target setting parameters of the vehicle function under the target data corresponding to the target variables.

[0233] For example, the data that the weather on July 21 is sunny, the weather on July 22 is heavy rain, and the weather on July 23 is cloudy is input into the target travel model, and it can be predicted that the target travel time on July 21 is 9:00, and the target setting parameters are playing music and adjusting the air-conditioning temperature to 22°C; the target travel time on July 22 is 9:00, and the target setting parameters are playing music; the target travel time on July 23 is 9:00, and the target setting parameters are playing music and opening the windows by 20%.

[0234] In one implementation, since there are various variables that affect travel data, in order to improve the prediction accuracy of target travel data, a plurality of target variables with a high degree of influence on the travel data may be determined through a target second optimization algorithm.

[0235] Furthermore, for multiple target variables, different weight values ​​are assigned to them according to their influence on the travel data. In the process of predicting the target travel data, target variables with different weight values ​​are combined to further improve the prediction accuracy of the target travel data.

[0236] Exemplarily, the determined target variables include weather, date attributes, and seasons; date attributes include weekends, holidays, weekdays, etc.; based on the degree of influence, the weight value of the date attribute is determined to be 50%, the weight value of the weather is 30%, and the weight value of the season is 20%; in the prediction process, first determine the candidate travel data based on the target data corresponding to the date attributes; then optimize the candidate travel data based on the weather and season to obtain the target travel data.

[0237] Optionally, the optimization process of the above-mentioned second optimization algorithm and the training process of the target travel model can be found in Figure 2 and Figure 3 The related description of the embodiments of the present application will not be repeated here.

[0238] S403, based on the target setting parameters, controlling the vehicle functions at a preset time before the target travel time.

[0239] For example, according to the target data corresponding to the target variable on July 21, the target travel time on July 21 is predicted to be 9:00, and the target setting parameters are playing music and adjusting the air-conditioning temperature to 22°C; the air-conditioning temperature can be adjusted to 22°C in advance at 8:55, and prepare to play music, saving travel time for users.

[0240] In one implementation, in order to avoid invalid adjustments to vehicle functions, a prompt message can be sent to the user at a preset time before the target travel time, so that the user can determine whether it is necessary to adjust the vehicle functions in advance according to the target setting parameters; if the user confirms the adjustment, the vehicle functions are controlled; if the user refuses the adjustment, the vehicle functions are not controlled to meet the actual needs of the user.

[0241] In an embodiment of the present application, the target travel data of the vehicle is predicted by the trained target travel model, and the vehicle functions are controlled in advance according to the predicted target travel time and the target setting parameters of the vehicle functions, thereby improving the intelligence of the vehicle, saving the user's travel time, and improving the user's travel experience.

[0242] Figure 5 is a schematic flow chart of another method for predicting travel data provided in an embodiment of the present application. It should be understood that the method can be applied to vehicles.

[0243] For example, Figure 5 As shown, the method 500 includes:

[0244] S501, optimizing the second optimization algorithm by using the first optimization algorithm and t distribution to obtain a target second optimization algorithm.

[0245] Exemplarily, the second optimization algorithm is the Eagle optimization algorithm. The first optimization algorithm, such as the whale optimization algorithm, is introduced into the first search stage of the traditional Eagle optimization algorithm, so that the optimal variables can be searched within a larger range to avoid the variables falling into local optimality; after the variables are updated through the search stage, the updated variables are mutated and disturbed again through the t distribution to achieve a balance between global search and local search, thereby improving the processing performance and efficiency of the target second optimization algorithm in the process of determining the target variables.

[0246] Optionally, the implementation of S501 can refer to Figure 2 The relevant description of S202 and S203 in the embodiment of the present application is not repeated here.

[0247] S502, based on the historical travel data and the historical data corresponding to the historical variables, determine the target variable through the target second optimization algorithm.

[0248] Exemplarily, historical travel data of the vehicle and historical data corresponding to historical variables are obtained, and the target variable having a greater impact on the travel data is determined by optimizing the target second optimization algorithm.

[0249] Optionally, the implementation of S502 can refer to Figure 2 The relevant description of S203 in the present application embodiment will not be repeated here.

[0250] S503, training a target travel model based on the target variable and historical travel data.

[0251] Exemplarily, after determining the target variable that has a greater impact on the travel data, the travel model to be trained is trained based on historical data related to the target variable and historical travel data to obtain a trained target travel model, so that the target travel model has the function of predicting the target travel data.

[0252] Optionally, the implementation of S503 can refer to Figure 2 The relevant description of S203 in the present application embodiment will not be repeated here.

[0253] S504, obtaining target data corresponding to the target variable.

[0254] Exemplarily, during the application process, target data corresponding to the target variable in the future time period is obtained. For example, the target data corresponding to the weather variable on July 21 is sunny, and the target data corresponding to the weather variable on July 22 is heavy rain, etc.

[0255] Optionally, the implementation of S504 can refer to Figure 4 The relevant description of S401 in the present application embodiment is not repeated here.

[0256] S505, inputting the target data into the target travel model, and predicting the target travel time and the target setting parameters of the vehicle function.

[0257] Exemplarily, the acquired target data is input into a trained target travel model, and the target travel model predicts the target travel time of the vehicle and target setting parameters for setting vehicle functions at the target travel time based on the target data.

[0258] Optionally, the implementation of S505 can refer to Figure 4 The relevant description of S402 in the present application embodiment is not repeated here.

[0259] S506, based on the target setting parameters, control the vehicle functions at a preset time before the target travel time.

[0260] For example, if the predicted target travel time is 9:00, the vehicle functions can be controlled in advance at 8:55 according to the target setting parameters to save the user's travel time.

[0261] Optionally, the implementation of S506 can refer to Figure 4 The relevant description of S403 in the present application embodiment is not repeated here.

[0262] In the embodiment of the present application, the whale optimization algorithm and the t-distribution optimization second optimization algorithm are used to obtain the target second optimization algorithm, thereby improving the global optimization capability of the algorithm; the target variable is determined by the target second optimization algorithm, thereby improving the efficiency and accuracy of searching for the target variable; the target variable and historical travel data are combined to train a target travel model, thereby enabling the target travel model to have the ability to predict target travel data; during the application process, the target travel data is predicted by the target travel model, thereby improving the accuracy of the prediction; and the vehicle functions are further controlled in advance based on the target travel data, thereby saving time for users and improving their travel experience.

[0263] Combination of the above Figures 1 to 5 A method for training a travel model and a method for predicting travel data provided by an embodiment of the present application are described in detail; Figures 6 to 8 The device embodiments of the present application are described in detail. It should be understood that the device in the embodiments of the present application can execute the various methods of the aforementioned embodiments of the present application, that is, the specific working processes of the following various products can refer to the corresponding processes in the aforementioned method embodiments.

[0264] Figure 6 It is a structural schematic diagram of a travel model training device provided in an embodiment of the present application.

[0265] For example, Figure 6 As shown, the device 600 includes:

[0266] The first acquisition module 601 is used to acquire training data; wherein the training data includes historical travel data of the vehicle and historical data corresponding to historical variables; the historical travel data includes historical travel time and historical setting parameters of vehicle functions; the historical variables indicate variables that have an impact on the historical travel data;

[0267] The optimization module 602 is used to optimize the second optimization algorithm based on the first optimization algorithm and the t distribution to obtain a target second optimization algorithm; wherein the first optimization algorithm and the second optimization algorithm are used to represent different heuristic optimization algorithms; the target second optimization algorithm is used to determine the target variable according to the historical variable;

[0268] The training module 603 is used to train the travel model to be trained based on the target second optimization algorithm and the training data to obtain a trained target travel model; wherein the target travel model is used to predict the target travel data of the vehicle; the target travel data includes the target travel time and the target setting parameters of the vehicle function.

[0269] In one possible implementation, the second optimization algorithm includes the first search strategy, and the optimization module 602 is further used to optimize the first search strategy through the first optimization algorithm to obtain an optimized first search strategy; based on the optimized first search strategy and t distribution, a target second optimization algorithm is obtained.

[0270] In one possible implementation, the training module 603 is also used to iteratively process historical variables through a target second optimization algorithm to obtain a target variable; based on the target variable and historical travel data, the travel model to be trained is trained to obtain a trained target travel model.

[0271] In one possible implementation, the training module 603 is also used to obtain the current iteration number i and the variables corresponding to the i-th iteration; wherein i is a positive integer; when i is 1, the variables corresponding to the i-th iteration are historical variables; based on the current iteration number i and the target iteration number, the target search strategy is determined in the target second optimization algorithm; through the target search strategy and t-distribution, the variables corresponding to the i-th iteration are updated to obtain the updated variables corresponding to the i+1-th iteration; based on the updated variables corresponding to the i+1-th iteration, the target variables are determined.

[0272] In one possible implementation, the second optimization algorithm also includes a second search strategy, a third search strategy and a fourth search strategy; the training module 603 is also used to determine the second search strategy, or the third search strategy as the target search strategy if the current number of iterations i is less than or equal to the preset number; if the current number of iterations i is greater than the preset number, the optimized first search strategy, or the fourth search strategy is determined as the target search strategy.

[0273] In one possible implementation, the training module 603 is also used to obtain a first random number and a first target numerical range; wherein the first random number belongs to the first target numerical range; based on the first target numerical range, a first target threshold is determined; if the first random number is less than the first target threshold, the second search strategy is determined as the target search strategy; if the first random number is greater than or equal to the first target threshold, the third search strategy is determined as the target search strategy.

[0274] In one possible implementation, the training module 603 is also used to obtain a second random number and a second target numerical range; wherein the second random number belongs to the second target numerical range; based on the second target numerical range, a second target threshold is determined; if the second random number is less than the second target threshold, the optimized first search strategy is determined as the target search strategy; if the second random number is greater than or equal to the second target threshold, the fourth search strategy is determined as the target search strategy.

[0275] In one possible implementation, the training module 603 is also used to update the variable corresponding to the i-th iteration through a target search strategy to obtain a first candidate variable; determine a first fitness value of the variable corresponding to the i-th iteration and a second fitness value of the first candidate variable; determine a second candidate variable based on the first fitness value and the second fitness value; and update the second candidate variable through a t distribution to obtain an updated variable corresponding to the i+1-th iteration.

[0276] In one possible implementation, the training module 603 is also used to determine the variable corresponding to the i+1th iteration as the target variable if the current iteration number i is greater than the target iteration number; and / or, if the fitness value of the variable corresponding to the i+1th iteration reaches a preset fitness threshold, determine the variable corresponding to the i+1th iteration as the target variable.

[0277] Figure 7 It is a structural diagram of a travel data prediction device provided in an embodiment of the present application.

[0278] For example, Figure 7 As shown, the device 700 includes:

[0279] The second acquisition module 701 is used to acquire target data corresponding to the target variable;

[0280] Prediction module 702, used to input target data into the target travel model to obtain target travel data of the vehicle; wherein the target travel data includes a target travel time and target setting parameters for setting vehicle functions at the target travel time;

[0281] The control module 703 is used to control the vehicle functions at a preset time before the target travel time based on the target setting parameters; wherein, the training method of the target travel model refers to the content described in the above application embodiment.

[0282] It should be noted that the above vehicle display device is embodied in the form of a functional unit. The term "module" here can be implemented in the form of software and / or hardware, and is not specifically limited to this.

[0283] For example, a "module" may be a software program, a hardware circuit, or a combination of the two that implements the above functions. The hardware circuit may include an application specific integrated circuit (ASIC), an electronic circuit, a processor (such as a shared processor, a dedicated processor, or a group processor, etc.) and a memory for executing one or more software or firmware programs, a combined logic circuit, and / or other suitable components that support the described functions.

[0284] Therefore, the units of each example described in the embodiments of the present application can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present application.

[0285] Figure 8 It is a structural schematic diagram of a vehicle provided in an embodiment of the present application.

[0286] For example, Figure 8 As shown, the vehicle 800 includes: a memory 801 and a processor 802, wherein the memory 801 stores an executable program code 803, and the processor 802 is used to call and execute the executable program code 803 to perform a travel model training method.

[0287] In addition, an embodiment of the present application also protects a device, which may include a memory and a processor, wherein the memory stores executable program code, and the processor is used to call and execute the executable program code to execute a travel model training method and a travel data prediction method provided in an embodiment of the present application.

[0288] In this embodiment, the functional modules of the device can be divided according to the above method example. For example, each functional module can be corresponded, or two or more functions can be integrated into one processing module. The above integrated module can be implemented in the form of hardware. It should be noted that the division of modules in this embodiment is schematic and is only a logical function division. There may be other division methods in actual implementation.

[0289] In the case of dividing each functional module according to each function, the device may also include a first acquisition module, an optimization module, a training module, a second acquisition module, a prediction module, and a control module, etc. It should be noted that all relevant contents of each step involved in the above method embodiment can be referred to the functional description of the corresponding functional module, which will not be repeated here.

[0290] It should be understood that the device provided in this embodiment is used to execute the above-mentioned training method of a travel model and a travel data prediction method, and thus can achieve the same effect as the above-mentioned implementation method.

[0291] In the case of an integrated unit, the device may include a processing module and a storage module. When the device is applied to a vehicle, the processing module may be used to control and manage the actions of the vehicle. The storage module may be used to support the vehicle in executing mutual program codes, etc.

[0292] The processing module may be a processor or a controller, which may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. The processor may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of digital signal processing (DSP) and a microprocessor, etc. The storage module may be a memory.

[0293] In addition, the device provided in the embodiments of the present application may specifically be a chip, a component or a module, and the chip may include a connected processor and a memory; wherein the memory is used to store instructions, and when the processor calls and executes the instructions, the chip can execute a travel model training method and a travel data prediction method provided in the above embodiments.

[0294] This embodiment also provides a computer-readable storage medium, which stores computer program code. When the computer program code runs on a computer, the computer executes the above-mentioned related method steps to implement a travel model training method and a travel data prediction method provided in the above-mentioned embodiment.

[0295] This embodiment also provides a computer program product. When the computer program product runs on a computer, it enables the computer to execute the above-mentioned related steps to implement a travel model training method and a travel data prediction method provided in the above-mentioned embodiment.

[0296] Among them, the device, computer-readable storage medium, computer program product or chip provided in this embodiment is used to execute the corresponding method provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding method provided above, and will not be repeated here.

[0297] Through the description of the above implementation methods, technical personnel in the relevant field can understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0298] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of modules or units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0299] The above contents are only specific implementation methods of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for training a travel model, characterized in that: The method comprises: Acquire training data; wherein the training data includes historical travel data of the vehicle and historical data corresponding to historical variables; the historical travel data includes historical travel time and historical setting parameters of vehicle functions; the historical variables indicate variables that have an impact on the historical travel data; The second optimization algorithm is optimized based on the first optimization algorithm and the t distribution to obtain a target second optimization algorithm; wherein the first optimization algorithm and the second optimization algorithm are used to represent different heuristic optimization algorithms; the target second optimization algorithm is used to determine the target variable according to the historical variable; Based on the target second optimization algorithm and the training data, the travel model to be trained is trained to obtain a trained target travel model; wherein the target travel model is used to predict the target travel data of the vehicle; the target travel data includes the target travel time and the target setting parameters of the vehicle function.

2. The method according to claim 1, characterized in that The second optimization algorithm includes a first search strategy, and the second optimization algorithm is optimized based on the first optimization algorithm and t distribution to obtain a target second optimization algorithm, including: Optimizing the first search strategy by using the first optimization algorithm to obtain an optimized first search strategy; Based on the optimized first search strategy and the t-distribution, the target second optimization algorithm is obtained.

3. The method according to claim 2, characterized in that The step of training the travel model to be trained based on the target second optimization algorithm and the training data to obtain the trained target travel model includes: Iteratively process the historical variables through the target second optimization algorithm to obtain the target variables; Based on the target variable and the historical travel data, the travel model to be trained is trained to obtain the trained target travel model.

4. The method according to claim 3, characterized in that The iterative processing of the historical variables by the target second optimization algorithm to obtain the target variable includes: Get the current iteration number i and the variable corresponding to the i-th iteration; where i is a positive integer; when i is 1, the variable corresponding to the i-th iteration is the historical variable; Determining a target search strategy in the target second optimization algorithm based on the current iteration number i and a preset number; By using the target search strategy and the t-distribution, the variable corresponding to the i-th iteration is updated to obtain the updated variable corresponding to the i+1-th iteration; The target variable is determined based on the updated variable corresponding to the (i+1)th iteration.

5. The method according to claim 4, characterized in that The second optimization algorithm also includes a second search strategy, a third search strategy, and a fourth search strategy; The determining of the target search strategy in the target second optimization algorithm based on the current iteration number i and the preset number includes: If the current number of iterations i is less than or equal to the preset number, determining the second search strategy or the third search strategy as the target search strategy; If the current iteration number i is greater than the preset number, the optimized first search strategy or the fourth search strategy is determined as the target search strategy.

6. The method according to claim 5, characterized in that The step of determining the second search strategy or the third search strategy as the target search strategy includes: Obtaining a first random number and a first target numerical range; wherein the first random number belongs to the first target numerical range; Determining a first target threshold based on the first target value range; If the first random number is less than the first target threshold, determining the second search strategy as the target search strategy; If the first random number is greater than or equal to the first target threshold, the third search strategy is determined as the target search strategy.

7. The method according to claim 5, characterized in that The step of determining the optimized first search strategy or the fourth search strategy as the target search strategy includes: Obtaining a second random number and a second target numerical range; wherein the second random number belongs to the second target numerical range; Based on the second target value range, determining a second target threshold; If the second random number is less than the second target threshold, determining the optimized first search strategy as the target search strategy; If the second random number is greater than or equal to the second target threshold, the fourth search strategy is determined as the target search strategy.

8. The method according to any one of claims 4 to 7, characterized in that The updating of the variable corresponding to the i-th iteration by the target search strategy and the t-distribution to obtain the updated variable corresponding to the i+1-th iteration includes: The variable corresponding to the i-th iteration is updated by the target search strategy to obtain a first candidate variable; Determine a first fitness value of the variable corresponding to the i-th iteration and a second fitness value of the first candidate variable; Determining a second candidate variable based on the first fitness value and the second fitness value; The second candidate variable is updated by using the t distribution to obtain an updated variable corresponding to the (i+1)th iteration.

9. A method for predicting travel data, characterized in that: The method comprises: Get the target data corresponding to the target variable; Inputting the target data into a target travel model to obtain target travel data of the vehicle; wherein the target travel data includes a target travel time and target setting parameters for setting vehicle functions at the target travel time; Based on the target setting parameters, controlling the vehicle function at a preset time before the target travel time; Wherein, the training method of the target travel model is the method described in any one of claims 1 to 8.

10. A vehicle, characterized in that: The vehicle comprises: A memory for storing executable program codes; A processor, configured to call and run the executable program code from the memory so that the vehicle executes the method as claimed in any one of claims 1 to 8 or claim 9.

Images