Shared vehicle pre-scheduling method and system based on user data

Abstract

The invention relates to the technical field of shared vehicle scheduling, and discloses a shared vehicle pre-scheduling method and system based on user data, and the method comprises the steps: collecting scheduling data, carrying out the preprocessing and analysis of the collected scheduling data, building a vehicle space-time monitoring model through the preprocessed scheduling data, and carrying out the pre-scheduling of a shared vehicle. The method comprises the following steps: analyzing time and geography dependence of shared vehicle scheduling through a vehicle space-time monitoring model, establishing a user behavior model through preprocessed scheduling data, monitoring and predicting shared vehicle usage by a user, establishing a weather prediction model to monitor and predict vehicle scheduling environment conditions, and establishing a full attribute set; according to the method, the willingness of the user to use the shared vehicle is analyzed, the shared vehicle is pre-scheduled, a visual interface and a user interaction interface are established to bring environmental factors into a vehicle scheduling decision, and combined scheduling of weather and the shared vehicle is facilitated.

Description

Technical Field

[0001] The invention relates to the technical field of shared vehicle scheduling, and discloses a shared vehicle pre-scheduling method and system based on user data. Background Art

[0002] With the rapid development of shared vehicles, the demand for shared vehicle scheduling is also increasing. At present, there are still many problems in shared vehicle scheduling. For example, the accuracy of user demand forecasting is not high enough, especially the impact of factors such as emergencies and weather changes is insufficient. There are also errors in vehicle supply forecasting. For example, factors such as vehicle failures and traffic congestion are difficult to accurately predict.

[0003] When processing user data in the face of large-scale real-time data, the computing speed and efficiency of existing algorithms still have room for improvement. The algorithm needs to be further optimized to cope with large-scale scheduling needs during peak periods. The response speed to real-time situations is not fast enough, and it is difficult to adjust the pre-scheduling plan in time. There is a lack of flexible response mechanisms to various emergencies. The computational complexity in large-scale scenarios is high, and it is difficult to respond in real time and find the optimal solution within a limited time. Approximate or heuristic algorithms are often required. The algorithm is not scalable enough to cope with rapid growth in scale. The response to real-time changes is not sensitive enough, and it is difficult to quickly adjust the established scheduling plan. There is a lack of effective feedback mechanism to continuously optimize the scheduling strategy. Therefore, a new shared vehicle pre-scheduling method is urgently needed to help meet the needs of users for shared vehicles.

[0004] For example, a Chinese patent with publication number CN114358563A discloses a method, system, device and storage medium for intelligent scheduling of shared vehicles; the intelligent scheduling method for shared vehicles includes: obtaining an image of the parking area of ​​the shared vehicle; the parking area includes a parking area and / or a prohibited parking area; parsing the image to identify the current parking information of the shared vehicle in the parking area; based on the current parking information and the preset parking information of the parking area, generating the scheduling information of the shared vehicle in the parking area. This method greatly improves the scheduling efficiency of shared vehicle parking, avoids the problem of shared vehicles being piled up in a centralized manner or not being able to find shared vehicles in some areas, and improves the utilization rate of shared vehicles.

[0005] The image quality of the above-mentioned patent is affected by factors such as weather and light, which may affect the recognition accuracy. Large-scale real-time image processing requires powerful computing resources, which makes it difficult to capture dynamic changes and may have lags. The preset information may not be able to reflect changes in actual conditions in a timely manner, and it is difficult to respond to emergencies or temporarily changed parking rules. It ignores changes in demand caused by seasonal or special events. The entire process from image acquisition to scheduling information generation may take a long time and is difficult to respond quickly to sudden large-scale demands, resulting in a lag with users' real-time needs. Summary of the invention

[0006] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0007] In order to solve the above technical problems, the main purpose of the present invention is to provide a shared vehicle pre-scheduling method based on user data, comprising:

[0008] S1. Collect scheduling data, and pre-process and analyze the collected scheduling data;

[0009] S2. Establish a vehicle spatiotemporal monitoring model through the preprocessed dispatch data, and analyze the time and geographical dependencies of the dispatch of shared vehicles through the vehicle spatiotemporal monitoring model;

[0010] S3. Establish a user behavior model through the pre-processed dispatch data to monitor and predict the user's use of shared vehicles;

[0011] S4. Establish a weather forecast model to monitor and predict the vehicle dispatch environment, establish a full attribute set, analyze the user's willingness to use shared vehicles, and pre-dispatch shared vehicles;

[0012] S5. Establish visualization interface and user interaction interface.

[0013] As a preferred solution of the shared vehicle pre-scheduling method based on user data of the present invention, wherein:

[0014] Taking time characteristics and geographic characteristics as basic data, a vehicle spatiotemporal monitoring model is established. The calculation expression of the vehicle spatiotemporal monitoring model is as follows:

[0015] F1(x)=f[f(ω1x+b1)ω2+b2]

[0016] Among them, F1(x) is the time and geographical dependency of the vehicle scheduling output by the vehicle spatiotemporal monitoring model, ω1 is the time feature, ω2 is the geographical feature, x is the input shared vehicle historical scheduling data, b1 is the time deviation value, b2 is the geographical data deviation value, and f() is the activation function.

[0017] As a preferred solution of the shared vehicle pre-scheduling method based on user data of the present invention, wherein:

[0018] The user behavior model calculates the negative gradient of the user's use of shared vehicles through user characteristics, and then outputs the probability of the user using shared vehicles;

[0019] The negative gradient calculation expression for users using shared vehicles is:

[0020]

[0021] Among them, γ is the negative gradient value, L(y i ,g(y i )) is the loss term between the predicted value and the actual value of user behavior, For mathematical partial derivative symbols, y i is the input user behavior data, g(y i ) is the user behavior model;

[0022] The user behavior model inputs user features through the initial user behavior model, iteratively updates the training set, and outputs the updated probability of the user using the shared vehicle. The user behavior model expression is as follows:

[0023]

[0024] Among them, ω3 is the user feature, y i is the user behavior data, b3 is the user behavior deviation, and n is the number of user behavior data groups;

[0025] The loss term expression is as follows:

[0026]

[0027]

[0028] Among them, y is the real-time data of user behavior, g(y) is the predicted data of user behavior model, and λ is the quantile.

[0029] As a preferred solution of the shared vehicle pre-scheduling method based on user data of the present invention, wherein:

[0030] Transform environmental data into time series attributes, establish weather forecast models, and monitor and predict the weather;

[0031] The weather forecast model calculation expression is as follows:

[0032]

[0033] Among them, Q is the weather forecast level, q is the weather data index, and q k is the weather data at the kth moment, i' k is the theoretical temperature data at the kth moment, i k is the actual temperature data at the kth moment, k is a sequence number, which can be 1, 2, 3, ..., N, and N is the total number of moments, which can be any positive integer;

[0034] The mean square error is used to calculate the difference between the predicted value and the actual value to evaluate the accuracy of the model.

[0035] As a preferred solution of the shared vehicle pre-scheduling method based on user data of the present invention, wherein:

[0036] Establish a full attribute set to schedule shared vehicles. The calculation expression of the full attribute set is as follows:

[0037]

[0038] Among them, P m is the regional vehicle dispatch coefficient, γ1 is the negative gradient value of the user using the shared vehicle at the first moment, Q1 is the weather condition at the first moment, is the compensation value for the weather impact on the user’s use of shared vehicles, δ1 is the deviation compensation value, and v is the counting time series.

[0039] As a preferred solution of the shared vehicle pre-scheduling method based on user data of the present invention, wherein:

[0040] Visual interfaces include real-time dashboards and predictive analytics;

[0041] Real-time dashboards include vehicle distribution, demand heat maps, and key performance indicators;

[0042] Predictive analysis includes visualization of future demand and comparison of multiple scheduling options;

[0043] The user interaction interface includes human-computer intervention and adjustment, decision suggestion risk assessment and user feedback improvement.

[0044] A shared vehicle pre-dispatching system based on user data, wherein:

[0045] Data module, used to collect and pre-process the scheduling data;

[0046] Model management module, including vehicle dispatch prediction model, vehicle spatiotemporal monitoring model, user behavior model, weather prediction model and full attribute set;

[0047] An optimization module, used to optimize the models in the model management module;

[0048] Human-computer interaction module, including visualization screen and user interaction interface.

[0049] As a preferred solution of the shared vehicle pre-dispatching system based on user data of the present invention, wherein:

[0050] The vehicle dispatch prediction model is used to predict and dispatch vehicles;

[0051] The vehicle spatiotemporal monitoring model predicts vehicle demand in different time periods and regions by receiving temporal and geographical features;

[0052] The user behavior model inputs user characteristics through the initial user behavior model, continuously iterates and updates the training set, and outputs the updated probability of users using shared vehicles;

[0053] Weather prediction models are used to fit environmental data and predict and monitor the environment;

[0054] By establishing a full attribute set, shared vehicles are scheduled, and personalized scheduling suggestions are generated by combining users' historical behavior data and real-time context information.

[0055] As a preferred solution of the shared vehicle pre-dispatching system based on user data of the present invention, wherein:

[0056] The user behavior model calculates the negative gradient of users using shared vehicles through user characteristics, and then outputs the probability of users using shared vehicles. The training set is updated through the error rate of the previous round of iterations, and the probability of users using shared vehicles is calculated. By fitting the output value of the vehicle spatiotemporal monitoring model, the probability of users using shared vehicles as a whole is predicted, thereby pre-scheduling shared vehicles.

[0057] Beneficial effects of the present invention:

[0058] The present invention integrates multiple data sources and extracts complex features through deep feature engineering, which can capture information in more dimensions and improve the accuracy and comprehensiveness of predictions. It uses reinforcement learning algorithms to make scheduling decisions, and can adapt to complex and changing environments through continuous learning and optimization to improve long-term benefits. It incorporates environmental factors into scheduling decisions, which helps to achieve the combined scheduling of weather and shared vehicles. It applies personalized recommendation technology to vehicle scheduling, which can better meet the needs of different user groups and improve user satisfaction. It introduces real-time feedback and dynamic adjustment mechanisms to improve the response speed and adaptability of the system and can quickly respond to emergencies. The system has self-learning and optimization capabilities, can continuously improve performance over time, and adapt to market changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0060] Figure 1 This is a flow chart of a shared vehicle pre-scheduling method based on user data of the present invention;

[0061] Figure 2 This is a composition diagram of a shared vehicle pre-dispatching system based on user data of the present invention. DETAILED DESCRIPTION

[0062] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0063] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0064] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0065] Embodiment 1:

[0066] like Figure 1 As shown, a shared vehicle pre-scheduling method based on user data includes:

[0067] S1. Collect scheduling data, and pre-process and analyze the collected scheduling data;

[0068] The dispatch data includes vehicle data, user data and environmental data;

[0069] Vehicle data includes vehicle location, vehicle status and traffic data;

[0070] Vehicle location continuously records the vehicle's current location;

[0071] Vehicle status includes operating status, such as idle or in service;

[0072] Traffic data includes information such as traffic congestion and speed limits, which affect vehicle arrival times and route selection.

[0073] User data includes behavior data, habit data, and interaction data;

[0074] Behavioral data includes the time and specific location of users’ use of shared vehicles, the frequency of users’ use of vehicles and the duration of each use, and the distance of each trip, which can help understand users’ travel range and preferences;

[0075] Habit data includes recording the starting points and destinations that users frequently choose, and analyzing the time of day when users tend to use the vehicle;

[0076] Interaction data includes users’ interactive behaviors in mobile applications, such as search history, booking records, etc., as well as users’ feedback and evaluation of services, reflecting user satisfaction and needs.

[0077] Environmental data includes weather conditions and traffic conditions;

[0078] Weather conditions include weather conditions during the user's trip, such as temperature, precipitation, etc.;

[0079] Traffic conditions include traffic congestion or real-time traffic information during the user's journey.

[0080] Preprocessing includes removing outliers and duplicate data, unifying data formats, and handling missing values;

[0081] Outlier and Duplicate Data Removal is used to identify and remove outliers, such as unreasonable timestamps or location data, and duplicate records;

[0082] Unified data format ensures that all data is in a unified format for easy subsequent processing and analysis;

[0083] Handling missing values ​​is used to fill in missing values, such as using historical averages or interpolation methods.

[0084] Analyze the preprocessed scheduling data by establishing a feature engineering set;

[0085] Feature engineering includes time features, geographic features, and user features;

[0086] Time features include hours, dates, weeks, and seasons. Time features are used to reveal the periodic and seasonal changes in travel patterns.

[0087] Geographic features include hot spots and POI information. Geographic features are used to identify hot spots and important geographic landmarks (POIs) in the city. This information affects users' travel choices and vehicle distribution.

[0088] S2. Establish a vehicle spatiotemporal monitoring model through the preprocessed dispatch data, and analyze the time and geographical dependencies of the dispatch of shared vehicles through the vehicle spatiotemporal monitoring model. Establish a dispatch vehicle spatiotemporal monitoring model to capture the time and space dependencies of dispatch vehicles;

[0089] The vehicle spatiotemporal monitoring model is used to predict vehicle demand in different time periods and regions. By receiving time features and geographic features, the vehicle spatiotemporal monitoring model can learn vehicle usage patterns in different time periods and locations, thereby predicting future demand;

[0090] Taking time characteristics and geographic characteristics as basic data, a vehicle spatiotemporal monitoring model is established. The calculation expression of the vehicle spatiotemporal monitoring model is as follows:

[0091] F1(x)=f[f(ω1x+b1)ω2+b2]

[0092] Among them, F1(x) is the time and geographical dependency of the vehicle scheduling output by the vehicle spatiotemporal monitoring model, ω1 is the time feature, ω2 is the geographical feature, x is the input shared vehicle historical scheduling data, b1 is the time deviation value, b2 is the geographical data deviation value, and f() is the activation function;

[0093] The activation function is a built-in function used to map function results and activate data calculations.

[0094] S3. Establish a user behavior model through the pre-processed dispatch data to monitor and predict the user's use of shared vehicles;

[0095] A user behavior model is established based on user data and user characteristics. User preferences are predicted through collaborative filtering, and the probability of using shared vehicles is calculated based on user preferences, which helps predict the needs of specific users in different regions and time periods.

[0096] The user behavior model calculates the negative gradient of the user's use of shared vehicles through user characteristics, and then outputs the probability of the user using shared vehicles. The weight of the training set is updated by the error rate of the previous round of iterations. In this way, the probability of the user using shared vehicles can be calculated through the minimum error after rounds of iterations. The output value of the vehicle spatiotemporal monitoring model is fitted to predict the probability of the user using shared vehicles as a whole, so as to pre-dispatch shared vehicles.

[0097] The negative gradient calculation expression for users using shared vehicles is:

[0098]

[0099] Among them, γ is the negative gradient value, L(y i ,g(y i )) is the loss term between the predicted value and the actual value of user behavior, For mathematical partial derivative symbols, y i is the input user behavior data, g(y i ) is the user behavior model;

[0100] The user behavior model inputs user features through the initial user behavior model, continuously iterates and updates the training set, and outputs the updated probability of the user using the shared vehicle. The user behavior model expression is as follows:

[0101]

[0102] Among them, ω3 is the user feature, y i is the user behavior data, b3 is the user behavior deviation, and n is the number of user behavior data groups;

[0103] The loss term expression is as follows:

[0104]

[0105]

[0106] Where y is the real-time data of user behavior, g(y) is the predicted data of the user behavior model, and λ is the quantile, which takes the value of (0,1).

[0107] The loss term expression is calculated by fitting the real-time user behavior data and the predicted user behavior data, and then calculating the loss term predicted by the user behavior model. The negative gradient of the user's use of shared vehicles is calculated by taking the partial derivative of the user behavior model predicted data through the loss term.

[0108] Furthermore, the loss term needs to be continuously iterated. The minimum value is calculated by iteratively calculating the loss term of the i-th item's real-time user behavior data and the i-th item's predicted user behavior data, and the minimum value is assigned to the i+1-th item. The i+1-th loss term is substituted into the negative gradient calculation expression of the user using the shared vehicle, and then the i+1-th negative gradient is calculated. The probability of the user using the shared vehicle is calculated through the negative gradient value, and so on. It is continuously iterated, and finally the probability of the user using the shared vehicle is output through the minimum negative gradient fitted by the minimum error.

[0109] Establish multi-source data fusion and deep feature engineering, integrate multiple data sources, extract complex features through deep feature engineering, combine spatiotemporal series prediction and individual user behavior prediction, use reinforcement learning algorithms for scheduling decisions, and introduce real-time feedback and dynamic adjustment mechanisms.

[0110] S4. Establish a weather forecast model to monitor and predict the vehicle dispatch environment, establish a full attribute set, analyze the user's willingness to use shared vehicles, and pre-dispatch shared vehicles;

[0111] Transform environmental data into time series attributes, such as hourly average temperature, daily rainfall probability, wind speed change trend, etc., and combine historical data analysis to determine which weather attributes have the greatest impact on vehicle demand and usage patterns, such as increased vehicle demand in hot weather;

[0112] The vehicle location and weather data are received in real time through streaming processing. The weather prediction model is used to fit the environmental data and predict and monitor the environment. The weather prediction model fits the factors that affect the user's use of shared vehicles from two aspects: one is the weather condition and the other is the temperature condition. The calculation expression of the weather prediction model is as follows:

[0113]

[0114] Among them, Q is the weather forecast level, q is the weather data index, and q k is the weather data at the kth moment, i' k is the theoretical temperature data at the kth moment, i k is the actual temperature data at the kth moment, k is a sequence number, which can be 1, 2, 3, ..., N, and N is the total number of moments, which can be any positive integer;

[0115] Weather conditions include rainfall, snowfall, etc.;

[0116] The mean square error is used to calculate the difference between the predicted value and the actual value to evaluate the accuracy of the model.

[0117] Establish a full attribute set to schedule shared vehicles, combine the user's historical behavior data and real-time context information to generate personalized scheduling suggestions. The calculation expression of the full attribute set is as follows:

[0118]

[0119] Among them, P m is the regional vehicle dispatch coefficient, γ1 is the negative gradient value of the user using the shared vehicle at the first moment, Q1 is the weather condition at the first moment, Compensation for users who use shared vehicles due to weather conditions. The value is (0,1), δ1 is the deviation compensation value, and v is the counting timing.

[0120] It is expressed as the negative gradient of the user using the shared vehicle at the first moment is γ1, and the corresponding weather attribute is Q1. This set is multiplied by the compensation value of the user's use of the shared vehicle affected by the weather of Q1, and the vehicle scheduling coefficient under the influence of weather is output;

[0121] By calculating the maximum probability of the gradient of users using shared vehicles at a certain moment under the weather conditions at this moment, vehicles are dispatched for users including, for example, user groups that frequently use the vehicles in the morning and user groups that frequently use the vehicles at night.

[0122] Shared vehicle dispatch can better understand and respond to the needs of different user groups, and achieve differentiated and personalized service recommendations and dispatch strategies;

[0123] Through a variety of advanced artificial intelligence technologies, the accuracy and efficiency of shared vehicle pre-dispatch can be significantly improved. Through continuous learning and optimization, the system can continuously adapt to user needs and market changes to provide users with a better service experience; environmental factors and operating cost considerations are incorporated into dispatch decisions, and highly visual decision support tools are integrated to allow manual intervention and adjustment of automated decisions.

[0124] S5. Establish visualization interface and user interaction interface;

[0125] Visual interfaces include real-time dashboards and predictive analytics;

[0126] Real-time dashboards include vehicle distribution, demand heat maps, and key performance indicators;

[0127] The current vehicle distribution and the location of all available vehicles are marked on the map, and the status of the vehicle (free, reserved, in use, etc.) can be displayed by color or logo.

[0128] The demand heat map shows the demand heat in different areas based on user requests and historical data, helping to predict future demand and optimize vehicle allocation.

[0129] Key performance indicators include utilization rate and turnover rate, which show the average utilization rate and turnover rate of vehicles, help evaluate the efficiency and service level of the system, and show the average response time of the system to user requests, which is used to measure service efficiency.

[0130] Predictive analysis includes visualization of future demand and comparison of multiple scheduling options;

[0131] Future demand visualization uses time series analysis or forecasting models to display the forecast trend of vehicle demand in different time periods, providing intuitive charts and trend lines to help managers understand future demand changes and make corresponding adjustments.

[0132] Comparison of multiple scheduling schemes Based on the forecast data, different scheduling schemes are generated and compared and analyzed to visualize the advantages and disadvantages of different schemes, such as cost, service level, response time and other indicators, to help decision makers choose the optimal scheduling strategy.

[0133] The user interface includes human-machine intervention and adjustment, decision suggestion risk assessment and user feedback improvement;

[0134] Manual intervention and adjustment allow administrators to manually intervene in the system's vehicle allocation and scheduling decisions, and provide an intuitive operation interface, such as dragging vehicle icons or setting priorities, for flexible adjustments.

[0135] Decision recommendation risk assessment provides system-generated scheduling recommendations, such as recommended vehicle allocation plans or strategies to cope with forecasted demand, and evaluates the risks and impacts of various decisions based on real-time data and forecast results.

[0136] User feedback improvement is used to collect user feedback, understand user experience and needs, and improve the user interface and service strategies based on feedback to enhance user satisfaction and system efficiency.

[0137] Embodiment 2:

[0138] like Figure 2 As shown, a shared vehicle pre-dispatching system based on user data includes:

[0139] Data module, used to collect and pre-process the scheduling data;

[0140] Model management module, including vehicle dispatch prediction model, vehicle spatiotemporal monitoring model, user behavior model, weather prediction model and full attribute set;

[0141] The vehicle scheduling prediction model is used to predict and schedule vehicles. The vehicle scheduling prediction model uses the vehicle spatiotemporal monitoring model, user behavior model, weather prediction model and full attribute set as data basis to schedule vehicles. This application does not impose algorithm restrictions on the vehicle scheduling prediction model;

[0142] The vehicle spatiotemporal monitoring model predicts vehicle demand in different time periods and regions by receiving temporal and geographical features;

[0143] The user behavior model inputs user characteristics through the initial user behavior model, continuously iterates and updates the training set, and outputs the updated probability of users using shared vehicles;

[0144] Weather prediction models are used to fit environmental data and predict and monitor the environment;

[0145] By establishing a full attribute set, shared vehicles are dispatched, and personalized dispatch suggestions are generated by combining users' historical behavior data and real-time context information;

[0146] An optimization module, used to optimize the models in the model management module;

[0147] Human-computer interaction module, including visualization screen and user interaction interface.

[0148] Importantly, it should be noted that the construction and arrangement of the present application shown in a plurality of different exemplary embodiments are only exemplary. Although only two embodiments are described in detail in this disclosure, it should be readily understood by those who refer to this disclosure that many modifications are possible, for example, the size, scale, structure, shape and ratio of various elements, and parameter values ​​(e.g., temperature, pressure, etc.), installation arrangement, use of materials, color, directional changes, etc., without substantially departing from the novel teachings and advantages of the subject matter described in the application. For example, the element shown as integrally formed can be composed of multiple parts or elements, the position of the element can be inverted or otherwise changed, and the nature or number or position of the discrete element can be changed or changed. Therefore, all such modifications are intended to be included in the scope of the present invention. The order or sequence of any process or method step can be changed or reordered according to an alternative embodiment. Any "device plus function" clause is intended to cover the structure of the execution function described in this article, and is not only structurally equivalent but also equivalent structure. Without departing from the scope of the present invention, other replacements, modifications, changes and omissions can be made in the design, operating conditions and arrangement of the exemplary embodiment. Therefore, the invention is not limited to a specific embodiment, but extends to numerous modifications still falling within the scope of the appended claims.

[0149] Additionally, in order to provide a concise description of exemplary embodiments, all features of an actual embodiment (ie, those features that are not relevant to the best mode presently contemplated for carrying out the invention or those that are not relevant to implementing the invention) may not be described.

[0150] It should be understood that in the development of any actual implementation, as in any engineering or design project, numerous implementation-specific decisions may be made. Such a development effort may be complex and time-consuming, but for those of ordinary skill having the benefit of this disclosure, the development effort will be a routine task of design, fabrication, and production without undue experimentation.

[0151] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A shared vehicle pre-scheduling method based on user data, characterized in that: include: S1. Collect scheduling data, and pre-process and analyze the collected scheduling data; S2. Establish a vehicle spatiotemporal monitoring model through the preprocessed dispatch data, and analyze the time and geographical dependencies of the dispatch of shared vehicles through the vehicle spatiotemporal monitoring model; S3. Establish a user behavior model through the pre-processed dispatch data to monitor and predict the user's use of shared vehicles; S4. Establish a weather forecast model to monitor and predict the vehicle dispatch environment, establish a full attribute set, analyze the user's willingness to use shared vehicles, and pre-dispatch shared vehicles; S5. Establish visualization interface and user interaction interface.

2. A shared vehicle pre-scheduling method based on user data according to claim 1, characterized in that: Taking time characteristics and geographic characteristics as basic data, a vehicle spatiotemporal monitoring model is established. The calculation expression of the vehicle spatiotemporal monitoring model is as follows: F1(x)=f[f(ω1x+b1)ω2+b2] Among them, F1(x) is the time and geographical dependency of the vehicle scheduling output by the vehicle spatiotemporal monitoring model, ω1 is the time feature, ω2 is the geographical feature, x is the input shared vehicle historical scheduling data, b1 is the time deviation value, b2 is the geographical data deviation value, and f() is the activation function.

3. A shared vehicle pre-scheduling method based on user data according to claim 2, characterized in that: The user behavior model calculates the negative gradient of the user's use of shared vehicles through user characteristics and outputs the probability of the user using shared vehicles; The negative gradient calculation expression for users using shared vehicles is: Among them, γ is the negative gradient value, L(y i ,g(y i )) is the loss term between the predicted value and the actual value of user behavior, For mathematical partial derivative symbols, y i is the input user behavior data, g(y i ) is the user behavior model; The user behavior model inputs user features through the initial user behavior model, iteratively updates the training set, and outputs the updated probability of the user using the shared vehicle. The user behavior model expression is as follows: Among them, ω3 is the user feature, y i is the user behavior data, b3 is the user behavior deviation, and n is the number of user behavior data groups; The loss term expression is as follows: Among them, y is the real-time data of user behavior, g(y) is the predicted data of user behavior model, and λ is the quantile.

4. A shared vehicle pre-scheduling method based on user data according to claim 3, characterized in that: Transform environmental data into time series attributes, establish weather forecast models, and monitor and predict the weather; The weather forecast model calculation expression is as follows: Among them, Q is the weather forecast level, q is the weather data index, and q k is the weather data at the kth moment, i′ k is the theoretical temperature data at the kth moment, i k is the actual temperature data at the kth moment, k is a sequence number, which can be 1, 2, 3, ..., N, and N is the total number of moments, which can be any positive integer; The mean square error is used to calculate the difference between the predicted value and the actual value to evaluate the accuracy of the model.

5. A shared vehicle pre-scheduling method based on user data according to claim 4, characterized in that: Establish a full attribute set to schedule shared vehicles. The calculation expression of the full attribute set is as follows: Among them, P m is the regional vehicle dispatch coefficient, γ1 is the negative gradient value of the user using the shared vehicle at the first moment, Q1 is the weather condition at the first moment, is the compensation value for the weather impact on the user’s use of shared vehicles, δ1 is the deviation compensation value, and v is the counting time series.

6. A shared vehicle pre-scheduling method based on user data according to claim 5, characterized in that: Visual interfaces include real-time dashboards and predictive analytics; Real-time dashboards include vehicle distribution, demand heat maps, and key performance indicators; Predictive analysis includes visualization of future demand and comparison of multiple scheduling options; The user interaction interface includes human-computer intervention and adjustment, decision suggestion risk assessment and user feedback improvement.

7. A shared vehicle pre-dispatching system based on user data, used to implement a shared vehicle pre-dispatching method based on user data as claimed in any one of claims 1 to 6, characterized in that: include, Data module, used to collect and pre-process the scheduling data; Model management module, including vehicle dispatch prediction model, vehicle spatiotemporal monitoring model, user behavior model, weather prediction model and full attribute set; An optimization module, used to optimize the models in the model management module; Human-computer interaction module, including visualization screen and user interaction interface.

8. A shared vehicle pre-dispatching system based on user data according to claim 7, characterized in that: The vehicle dispatch prediction model is used to predict and dispatch vehicles; The vehicle spatiotemporal monitoring model predicts vehicle demand in different time periods and regions by receiving temporal and geographical features; The user behavior model inputs user characteristics through the initial user behavior model, continuously iterates and updates the training set, and outputs the updated probability of users using shared vehicles; Weather prediction models are used to fit environmental data and predict and monitor the environment; By establishing a full attribute set, shared vehicles are scheduled, and personalized scheduling suggestions are generated by combining users' historical behavior data and real-time context information.

9. A shared vehicle pre-dispatching system based on user data according to claim 7, characterized in that: The user behavior model calculates the negative gradient of users using shared vehicles through user characteristics, and then outputs the probability of users using shared vehicles. The training set is updated through the error rate of the previous round of iterations, and the probability of users using shared vehicles is calculated. By fitting the output value of the vehicle spatiotemporal monitoring model, the probability of users using shared vehicles as a whole is predicted, thereby pre-scheduling shared vehicles.

Images