Method and device for overall planning and optimizing road rescue resource distribution
By analyzing vehicle road rescue data, training prediction models and optimizing resource distribution, the problems of waste and high operating costs of road rescue resources in the existing technology are solved, and efficient resource utilization and predictive optimization are achieved.
Patent Information
- Application Number
- CN202411015458.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-26
- Publication Date
- 2024-11-12
AI Technical Summary
The existing road rescue resource distribution methods are difficult to reduce operating costs while ensuring timely response, which can easily lead to resource waste.
By collecting and analyzing the vehicle's road rescue data, generating time series data derived variables, location information derived variables and weight indexes, training the daily rescue number prediction model and safety redundancy threshold model, and optimizing the distribution of road rescue resources.
Effectively predict future road rescue situations, provide reliable resource distribution suggestions, reduce resource waste, and reduce operational costs.
Smart Images

Figure CN118941036A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of big data application technology for Internet of Vehicles, and in particular to a method and device for comprehensively optimizing the distribution of road rescue resources. Background Art
[0002] With the development of Internet of Vehicles technology and big data technology applications, vehicle data is becoming increasingly rich, complete and comprehensive, providing more directions and possibilities for improving vehicle product availability and optimizing car owners' car experience. It also helps to promote product evolution, improve manufacturing quality and efficiency, and enhance product competitiveness.
[0003] Regardless of new energy vehicles or fuel vehicles, the availability of road rescue is one of the functions that must be guaranteed. However, the existing road rescue resource distribution method often causes waste of road rescue resources. It is difficult to reduce the operating cost of road rescue while ensuring timely response of road rescue, which needs to be solved urgently. Summary of the invention
[0004] The present application provides a method and device for comprehensively optimizing the distribution of road rescue resources to solve the problems that the existing road rescue resource distribution method often causes waste of road rescue resources and it is difficult to reduce the operating costs of road rescue while ensuring timely response to road rescue.
[0005] The first aspect of the present application provides a method for comprehensively optimizing the distribution of road rescue resources, which is applied to the offline training stage and includes the following steps: collecting road rescue data of multiple vehicles corresponding to the target vehicle model; generating time series data derived variables, position information derived variables and weight indexes corresponding to the road rescue data, wherein the weight index includes a weekly weight index and a daily weight index; based on the time series data derived variables, the position information derived variables and the weight index, training a preset daily rescue number prediction model and a safety redundancy threshold model, so as to optimize the distribution of road rescue resources through the trained daily rescue number prediction model and the safety redundancy threshold model.
[0006] Optionally, in one embodiment of the present application, the generation of time series data derived variables, location information derived variables and weight index corresponding to the road rescue data includes: obtaining first road rescue data of the multiple vehicles on each non-holiday in a preset first time period; obtaining second road rescue data of the multiple vehicles on each holiday in a preset second time period; generating the time series data derived variables and the location information derived variables based on the first road rescue data and the second road rescue data; based on a preset weekly summary strategy, summarizing all non-holidays in the preset first time period to obtain a non-holiday summary result, and calculating the weekly weight index based on the weekly summary result and the first road rescue data; based on a preset holiday summary strategy, summarizing all holidays in the preset second time period to obtain a holiday summary result, and calculating the daily weight index based on the holiday summary result and the second road rescue data.
[0007] Optionally, in one embodiment of the present application, the preset daily rescue times prediction model and safety redundancy threshold model are trained based on the time series data derived variables, the location information derived variables and the weight index to optimize the distribution of road rescue resources through the trained daily rescue times prediction model and safety redundancy threshold model, including: training the daily rescue times prediction model according to the time series data derived variables, the location information derived variables and the weight index to obtain the training prediction score of the daily rescue times prediction model; when the training prediction score is not less than the preset score threshold, obtaining the predicted value and actual value of the daily rescue times prediction model, and training the safety redundancy threshold model based on the predicted value and the actual value to generate a safety redundancy threshold; optimizing the distribution of road rescue resources through the predicted value and the safety redundancy threshold.
[0008] The second aspect of the present application provides a method for comprehensively optimizing the distribution of road rescue resources, which is applied to the online detection stage, and includes the following steps: obtaining target road rescue data corresponding to the target vehicle model; inputting the target road rescue data into a pre-trained daily rescue number prediction model to output the predicted rescue number in the target future time period, and obtaining the safety redundancy threshold corresponding to the predicted rescue number based on the predicted rescue number and the pre-trained safety redundancy threshold model; determining a road rescue resource distribution optimization suggestion based on the predicted rescue number and the safety redundancy threshold, so as to adjust the road rescue resource distribution through the road rescue resource distribution optimization suggestion.
[0009] Optionally, in one embodiment of the present application, after adjusting the road rescue resource distribution through the road rescue resource distribution optimization suggestion, it also includes: based on a preset detection cycle, detecting the pre-trained daily rescue number prediction model and the safety redundancy threshold model to obtain a corresponding KS indicator; judging whether the KS indicator meets a preset stability condition; if the KS indicator meets the preset stability condition, continuing to use the pre-trained daily rescue number prediction model and the safety redundancy threshold model to optimize the road rescue resource distribution; if the KS indicator does not meet the preset stability condition, retraining the daily rescue number prediction model and the safety redundancy threshold model.
[0010] The third aspect of the present application provides an apparatus for comprehensively optimizing the distribution of road rescue resources, which is applied to an offline training stage and includes: a first acquisition module for collecting road rescue data of multiple vehicles corresponding to a target vehicle model; a generation module for generating time series data derived variables, position information derived variables and weight indexes corresponding to the road rescue data, wherein the weight index includes a weekly weight index and a daily weight index; a training module for training a preset daily rescue times prediction model and a safety redundancy threshold model based on the time series data derived variables, the position information derived variables and the weight index, so as to optimize the distribution of road rescue resources through the trained daily rescue times prediction model and the safety redundancy threshold model.
[0011] Optionally, in one embodiment of the present application, the generation module includes: a first acquisition unit, used to acquire the first road rescue data of the multiple vehicles on each non-holiday within a preset first time period; a second acquisition unit, used to acquire the second road rescue data of the multiple vehicles on each holiday within a preset second time period; a third acquisition unit, used to generate the time series data derived variables and the location information derived variables based on the first road rescue data and the second road rescue data; a first aggregation unit, used to aggregate all non-holidays within the preset first time period based on a preset weekly aggregation strategy to obtain a non-holiday aggregation result, and calculate the weekly weight index based on the weekly aggregation result and the first road rescue data; a second aggregation unit, used to aggregate all holidays within the preset second time period based on a preset holiday aggregation strategy to obtain a holiday aggregation result, and calculate the daily weight index based on the holiday aggregation result and the second road rescue data.
[0012] Optionally, in one embodiment of the present application, the training module includes: a prediction unit, used to train the daily rescue times prediction model according to the time series data derived variables, the location information derived variables and the weight index to obtain the training prediction score of the daily rescue times prediction model; a fourth acquisition unit, used to obtain the predicted value and actual value of the daily rescue times prediction model when the training prediction score is not less than a preset score threshold, and train the safety redundancy threshold model based on the predicted value and the actual value to generate a safety redundancy threshold; an optimization unit, used to optimize the road rescue resource distribution through the predicted value and the safety redundancy threshold.
[0013] The fourth aspect of the present application provides a device for comprehensively optimizing the distribution of road rescue resources, which is applied to the online prediction stage, including: a second acquisition module, used to obtain target road rescue data corresponding to the target vehicle model; an input module, used to input the target road rescue data into a pre-trained daily rescue number prediction model to output the predicted rescue number in the target future time period, and obtain the safety redundancy threshold corresponding to the predicted rescue number based on the predicted rescue number and the pre-trained safety redundancy threshold model; an adjustment module, used to determine a road rescue resource distribution optimization suggestion based on the predicted rescue number and the safety redundancy threshold, so as to adjust the road rescue resource distribution through the road rescue resource distribution optimization suggestion.
[0014] Optionally, in one embodiment of the present application, it also includes: a detection module, which is used to detect the pre-trained daily rescue times prediction model and the safety redundancy threshold model based on a preset detection cycle after adjusting the road rescue resource distribution through the road rescue resource distribution optimization suggestion, so as to obtain a corresponding KS indicator; a judgment module, which is used to judge whether the KS indicator meets a preset stability condition; a stabilization module, which is used to continue to optimize the road rescue resource distribution using the pre-trained daily rescue times prediction model and the safety redundancy threshold model if the KS indicator meets the preset stability condition; and a loop module, which is used to retrain the daily rescue times prediction model and the safety redundancy threshold model if the KS indicator does not meet the preset stability condition.
[0015] The fifth aspect of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for comprehensively optimizing the distribution of road rescue resources as described in the above embodiments.
[0016] The sixth aspect of the present application provides a computer-readable storage medium, which stores a computer program. When the program is executed by a processor, it implements the above method for comprehensively optimizing the distribution of road rescue resources.
[0017] Therefore, the embodiments of the present application have the following beneficial effects:
[0018] The embodiments of the present application can collect road rescue data of multiple vehicles corresponding to the target vehicle model; generate time series data derived variables, location information derived variables and weight index corresponding to the road rescue data, wherein the weight index includes a weekly weight index and a daily weight index; based on the time series data derived variables, location information derived variables and weight index, train the preset daily rescue number prediction model and safety redundancy threshold model to optimize the distribution of road rescue resources through the trained daily rescue number prediction model and safety redundancy threshold model. The present application models and summarizes the road rescue situation from the perspective of time and space through road rescue historical data and driving conditions, so as to predict future road rescue conditions and provide reliable suggestions for the distribution of road rescue resources to adjust and optimize the distribution of road rescue resources. Thus, the existing road rescue resource distribution method often causes waste of road rescue resources and is difficult to reduce the operating cost of road rescue while ensuring timely response to road rescue.
[0019] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:
[0021] Figure 1 A flowchart of a method for comprehensively optimizing the distribution of road rescue resources in an offline training phase provided according to an embodiment of the present application;
[0022] Figure 2 A schematic diagram of the logical architecture of a method for comprehensively optimizing the distribution of road rescue resources provided in one embodiment of the present application;
[0023] Figure 3 A schematic diagram of the execution logic of a method for comprehensively optimizing the distribution of road rescue resources provided in one embodiment of the present application;
[0024] Figure 4 A schematic diagram of a road rescue resource data communication process provided by an embodiment of the present application;
[0025] Figure 5 A flowchart of a method for comprehensively optimizing the distribution of road rescue resources in an online prediction phase provided according to an embodiment of the present application;
[0026] Figure 6This is an example diagram of a device for optimizing the distribution of road rescue resources in an offline training phase according to an embodiment of the present application;
[0027] Figure 7 This is an example diagram of a device for optimizing the distribution of road rescue resources in an online prediction phase according to an embodiment of the present application;
[0028] Figure 8 A schematic diagram of the structure of a vehicle provided in an embodiment of the present application.
[0029] Among them, 10-a device for comprehensively optimizing the distribution of road rescue resources in the offline training stage, 20-a device for comprehensively optimizing the distribution of road rescue resources in the online prediction stage; 101-a first acquisition module, 102-a generation module, 103-a training module; 201-a second acquisition module, 202-an input module, 203-an adjustment module; 801-a memory, 802-a processor, 803-a communication interface. DETAILED DESCRIPTION
[0030] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.
[0031] The following describes the method and device for optimizing the distribution of road rescue resources in an embodiment of the present application with reference to the accompanying drawings. In view of the problems mentioned in the above background technology, the present application provides a method for optimizing the distribution of road rescue resources in an overall manner. In the method, by collecting road rescue data of multiple vehicles corresponding to the target vehicle model; generating time series data derived variables, location information derived variables and weight indexes corresponding to the road rescue data, wherein the weight index includes a weekly weight index and a daily weight index; based on the time series data derived variables, location information derived variables and weight indexes, training a preset daily rescue number prediction model and a safety redundancy threshold model, so as to optimize the distribution of road rescue resources through the trained daily rescue number prediction model and the safety redundancy threshold model. The present application models and summarizes the road rescue situation from the perspective of time and space through road rescue historical data and driving conditions, so as to predict future road rescue conditions, provide reliable suggestions for the distribution of road rescue resources, and adjust and optimize the distribution of road rescue resources. Thus, the existing road rescue resource distribution method often causes waste of road rescue resources, and it is difficult to reduce the operating cost of road rescue while ensuring timely response to road rescue.
[0032] Specifically, Figure 1A flowchart of a method for comprehensively optimizing the distribution of road rescue resources during an offline training phase provided in an embodiment of the present application.
[0033] like Figure 1 As shown, the method for comprehensively optimizing the distribution of road rescue resources includes the following steps:
[0034] In step S101 , roadside assistance data of multiple vehicles corresponding to the target vehicle type are collected.
[0035] In step S102, time series data derived variables, location information derived variables and weight indexes corresponding to the road rescue data are generated, wherein the weight indexes include a weekly weight index and a daily weight index.
[0036] The embodiments of the present application can first select batch (offline) road rescue data on regular dates in the past three months and batch (offline) road rescue data on holidays and before and after holidays in the past three years of multiple vehicles or vehicle-mounted transportation vehicles corresponding to the target vehicle model, so as to generate corresponding time series data derived variables, location information derived variables and weight indices such as weekly weight index and daily weight index.
[0037] It should be noted that in the embodiments of the present application, the above-mentioned vehicle or vehicle-mounted transportation vehicle is a vehicle driven by an engine to travel a distance in a broad sense, and in a narrow sense, includes new energy vehicles, sedans, SUVs, MPVs, commercial vehicles and the like. This vehicle needs to be equipped with a 2G / 3G / 4G vehicle-mounted communication module, which has the ability to exchange network communication and vehicle internal network signals, such as Figure 2 As shown, it can communicate with the back-end service platform used for vehicle connection management in the enterprise, that is, the Internet of Vehicles background.
[0038] Optionally, in one embodiment of the present application, time series data derived variables, location information derived variables and weight indexes corresponding to road rescue data are generated, including: obtaining first road rescue data of multiple vehicles on each non-holiday in a preset first time period; obtaining second road rescue data of multiple vehicles on each holiday in a preset second time period; generating time series data derived variables and location information derived variables based on the first road rescue data and the second road rescue data; based on a preset weekly aggregation strategy, aggregating all non-holidays in the preset first time period to obtain a non-holiday aggregation result, and calculating a weekly weight index based on the weekly aggregation result and the first road rescue data; based on a preset holiday aggregation strategy, aggregating all holidays in the preset second time period to obtain a holiday aggregation result, and calculating a daily weight index based on the holiday aggregation result and the second road rescue data.
[0039] In the actual implementation process, the embodiment of the present application can first obtain the GPS signal, occurrence time, and road rescue arrival time of the vehicle road rescue. The data must accurately reflect the actual status at the time, with clear information, and general data accumulation of ≥3 months, and special holiday data accumulation of ≥3 years, to ensure that the recognition output of the subsequent modeling can reflect the actual situation of road rescue;
[0040] In addition, the embodiments of the present application can generate derived variables such as the province, city, and district where the road rescue is located based on the GPS signal, that is, location information derived variables; and generate derived variables such as the day of the week when the road rescue occurs, whether it is a holiday, the beginning / middle / end of the month, and the number of rescues one / three / seven days ago based on the time when the road rescue occurs, that is, time series data derived variables; at the same time, based on the average daily number of rescues from Monday to Sunday and from the first day to the last day within three months, calculate the weekly weight index and daily weight index as a reference for the trend of the number of rescues over time.
[0041] Specifically, in the specific implementation process, the embodiments of the present application can be based on the number of rescues per day in the past three months, excluding holidays such as the Spring Festival, New Year's Day, and the related adjustment dates before and after, and summarized from Monday to Sunday to calculate the average daily number of rescues from Monday to Sunday; secondly, the minimum value is used as the weight of 1.0 to calculate the weight index from Monday to Sunday. The general cumulative result is between 7.0 and 14.0, which indirectly reflects the fluctuations between Monday and Sunday, and the predicted number of rescues for the remaining days is calculated through the number of rescues counted this week.
[0042] It should be noted that in the embodiments of the present application, the weekly weight index is often used in scenarios with a high correlation from Monday to Sunday, such as shopping mall clothing, marketing activities, etc. The embodiments of the present application mainly consider road rescue and vehicle travel, especially long-distance travel, which has a great relationship. Therefore, the weekly weight index needs to be selected as an important basic model and parameter.
[0043] In addition, the embodiments of the present application also need to calculate the average daily rescue number for each day from the first day to the Nth day of the holiday based on the number of rescues on holidays such as the Spring Festival and New Year's Day and the related adjustment dates before and after in the past three years; secondly, use the minimum value as the weight of 1.0 to calculate the weight index of other days; but unlike the weekly weight index, the daily weight index is mainly used to handle trend evaluation and prediction of non-routine dates such as promotional activities and holidays.
[0044] In step S103, based on the time series data derived variables, the location information derived variables and the weight index, the preset daily rescue times prediction model and the safety redundancy threshold model are trained to optimize the distribution of road rescue resources through the trained daily rescue times prediction model and the safety redundancy threshold model.
[0045] Furthermore, the embodiment of the present application generates corresponding derivative variables according to the GPS, occurrence time, etc. of the road rescue, such as Figure 3 As shown, a road rescue trend prediction model is established by combining the province, city, and district of road rescue, the number of weeks for road rescue, whether it is a holiday, and the weekly weight index and daily weight index, and mainly based on time series data processing; thereafter, the embodiment of the present application derives variables from the basic data, and trains the model separately according to whether it is related to holidays, so that it meets the fitting standard, thereby optimizing the distribution of road rescue resources.
[0046] Optionally, in one embodiment of the present application, a preset daily rescue number prediction model and a safety redundancy threshold model are trained based on time series data derived variables, location information derived variables and weight indexes to optimize the distribution of road rescue resources through the trained daily rescue number prediction model and safety redundancy threshold model, including: training the daily rescue number prediction model according to time series data derived variables, location information derived variables and weight indexes to obtain a training prediction score of the daily rescue number prediction model; when the training prediction score is not less than a preset score threshold, obtaining the predicted value and actual value of the daily rescue number prediction model, and training the safety redundancy threshold model based on the predicted value and the actual value to generate a safety redundancy threshold; optimizing the distribution of road rescue resources through the predicted value and the safety redundancy threshold.
[0047] In the process of specific matters, the embodiments of the present application can predict the number of future rescues based on the geographical location of road rescue (province, city, district), date attributes and time-series derived variables (number of rescues 1 / 3 / 7 days ago); the embodiments of the present application can use the Pearson coefficient to identify the correlation of model variables, so as to facilitate the screening and caliber interpretation of model variables, and use the random forest model for model training and generation. The model is enabled when the F1 score is greater than or equal to 70%, and the statistical and prediction results are summarized. It also provides road rescue demand forecasts by time (one day in a week or a month) and space (provinces, cities and districts in the country), thereby supporting the optimization of road rescue resource distribution.
[0048] Afterwards, the embodiments of the present application can statistically calculate the P99 (the probability that the actual value exceeds the predicted value + N is 1%) threshold based on the distribution of the predicted value and the actual value to determine the safety redundancy threshold, thereby providing downstream system resource distribution recommendations based on the predicted value and the safety redundancy threshold.
[0049] Those skilled in the art should understand that P99 is a method for assessing risk probability, which is mainly used in system operation and maintenance or manufacturing industries. It generally obeys a normal distribution and can be relatively stably distributed at both ends or one end. When ignoring low-probability events, it can effectively reduce the difference between event expectations and actual events.
[0050] To summarize, the embodiments of the present application generate geographic location and date-derived variables based on industry standards and data for subsequent data screening and judgment, and calculate weekly weight index and daily weight index to train the model to predict the number of future road rescues, and formulate safety redundancy thresholds based on the difference between the predicted results and the actual values, thereby effectively improving the accuracy and generalization of future road rescue situation predictions.
[0051] According to the method for optimizing the distribution of road rescue resources in the offline training stage proposed in the embodiment of the present application, the road rescue data of multiple vehicles corresponding to the target vehicle model are collected; time series data derived variables, location information derived variables and weight index corresponding to the road rescue data are generated, wherein the weight index includes a weekly weight index and a daily weight index; based on the time series data derived variables, location information derived variables and weight index, a preset daily rescue number prediction model and a safety redundancy threshold model are trained to optimize the distribution of road rescue resources through the trained daily rescue number prediction model and safety redundancy threshold model. The present application models and summarizes the road rescue situation from the perspective of time and space through road rescue historical data and driving conditions, etc., so as to predict future road rescue conditions and provide reliable suggestions for the distribution of road rescue resources to adjust and optimize the distribution of road rescue resources.
[0052] Figure 4 A flowchart of a method for comprehensively optimizing the distribution of road rescue resources in an online prediction stage provided in an embodiment of the present application.
[0053] like Figure 4 As shown, the method for comprehensively optimizing the distribution of road rescue resources includes the following steps:
[0054] In step S401, target roadside assistance data corresponding to the target vehicle type is obtained.
[0055] In step S402, the target road rescue data is input into a pre-trained daily rescue number prediction model to output the predicted rescue number in the target future time period, and the safety redundancy threshold corresponding to the predicted rescue number is obtained based on the predicted rescue number and the pre-trained safety redundancy threshold model.
[0056] In step S403, based on the predicted number of rescues and the safety redundancy threshold, a road rescue resource distribution optimization suggestion is determined to adjust the road rescue resource distribution through the road rescue resource distribution optimization suggestion.
[0057] After the daily rescue number prediction model is trained and made to meet the corresponding performance standards, the embodiment of the present application can migrate it to a big data real-time application cluster to perform batch road rescue resource distribution prediction.
[0058] Specifically, the embodiments of the present application can obtain road rescue data with the same parameters as the model, input it into the daily rescue number prediction model pre-trained in the cloud, and combine it with the safety redundancy threshold model to provide a safety redundancy resource allocation recommendation value of the predicted value according to the proofreading situation, that is, the possibility that the actual road rescue demand exceeds the predicted value by adding N% of resources is less than 1%, so as to generate the final real-time rescue work order (i.e., the road rescue resource distribution optimization recommendation), such as Figure 5 As shown, the distribution of road rescue resources is adjusted.
[0059] As an achievable method, the embodiments of the present application can output the estimated number of rescues by province, city, district and day, and pass it to the downstream system for corresponding processing. The road rescue system adjusts the rescue resource distribution in the next few days based on the predicted data. For example, the estimated number of road rescues in Jiangxi Province in the next week is 20 times, and the road rescue resources in Jiangxi Province are adjusted to support 30 times.
[0060] Therefore, the embodiments of the present application connect the statistical and predictive results to the downstream system, thereby displaying the statistical and predictive results so that it can timely coordinate and optimize the distribution of road rescue resources, greatly reducing the waste of resources.
[0061] Optionally, in one embodiment of the present application, after adjusting the road rescue resource distribution through road rescue resource distribution optimization suggestions, it also includes: based on a preset detection cycle, detecting the pre-trained daily rescue number prediction model and the safety redundancy threshold model to obtain the corresponding KS indicator; judging whether the KS indicator meets the preset stability condition; if the KS indicator meets the preset stability condition, continuing to use the pre-trained daily rescue number prediction model and the safety redundancy threshold model to optimize the road rescue resource distribution; if the KS indicator does not meet the preset stability condition, retraining the daily rescue number prediction model and the safety redundancy threshold model.
[0062] It is understandable that since the daily rescue number prediction model can only reflect the situation characteristics of the data used, especially for road rescue scenarios that will change over time, the embodiments of the present application need to continuously iterate the model. When the data accumulation meets the conditions, training iterations can be performed by time and geographical area.
[0063] It should be noted that the embodiments of the present application can test the daily rescue times prediction model every 3 months to obtain the corresponding KS index. When the KS value is <20%, it means that the model can no longer stably identify the corresponding scene, and the model parameters no longer have the ability to represent the scene logic. At this time, the embodiments of the present application need to use the data from the last 3 months for retraining and optimize the training parameters to ensure the quality and stability of the model.
[0064] Therefore, the embodiments of the present application can predict the number of rescues in time and space, which is more accurate and convenient, so as to coordinate and optimize the distribution of road rescue resources; at the same time, the road rescue results and predictions are combined to form a safety redundant design to reduce the risk of outliers; in addition, the embodiments of the present application improve the continuous iteration of data to construct or update the daily rescue number prediction model, so as to submit it to the downstream system as a reference for the overall optimization of the distribution of road rescue resources.
[0065] According to the method for optimizing the distribution of road rescue resources in the online prediction stage proposed in the embodiment of the present application, the target road rescue data corresponding to the target vehicle model is obtained; the target road rescue data is input into the pre-trained daily rescue number prediction model to output the predicted rescue number in the target future time period, and the safety redundancy threshold corresponding to the predicted rescue number is obtained based on the predicted rescue number and the pre-trained safety redundancy threshold model; based on the predicted rescue number and the safety redundancy threshold, the road rescue resource distribution optimization suggestion is determined to adjust the road rescue resource distribution through the road rescue resource distribution optimization suggestion. The present application models and summarizes the road rescue situation from the perspective of time and space through road rescue historical data and driving conditions, so as to predict future road rescue conditions and provide reliable suggestions for the distribution of road rescue resources to adjust and optimize the distribution of road rescue resources.
[0066] Secondly, a device for comprehensively optimizing the distribution of road rescue resources proposed in accordance with an embodiment of the present application is described with reference to the accompanying drawings.
[0067] Figure 6 It is a block diagram of an apparatus for comprehensively optimizing the distribution of road rescue resources in an offline training phase according to an embodiment of the present application.
[0068] like Figure 6 As shown, the device 10 for comprehensively optimizing the distribution of road rescue resources includes: a first acquisition module 101, a generation module 102 and a training module 103.
[0069] The first collection module 101 is used to collect roadside assistance data of multiple vehicles corresponding to the target vehicle model.
[0070] The generation module 102 is used to generate time series data derived variables, location information derived variables and weight indexes corresponding to the road rescue data, wherein the weight index includes a weekly weight index and a daily weight index.
[0071] The training module 103 is used to train a preset daily rescue number prediction model and a safety redundancy threshold model based on time series data derived variables, location information derived variables and weight indexes, so as to optimize the distribution of road rescue resources through the trained daily rescue number prediction model and safety redundancy threshold model.
[0072] Optionally, in one embodiment of the present application, the generation module 102 includes: a first acquisition unit, a second acquisition unit, a third acquisition unit, a first aggregation unit and a second aggregation unit.
[0073] The first acquisition unit is used to acquire first roadside assistance data of multiple vehicles on each non-holiday within a preset first time period.
[0074] The second acquisition unit is used to acquire second roadside assistance data of multiple vehicles on each holiday within a preset second time period.
[0075] The third acquisition unit is used to generate a time series data derived variable and a position information derived variable according to the first roadside assistance data and the second roadside assistance data.
[0076] The first aggregation unit is used to aggregate all non-holidays in a preset first time period based on a preset weekly aggregation strategy to obtain a non-holiday aggregation result, and calculate a weekly weight index based on the weekly aggregation result and the first road rescue data.
[0077] The second summarizing unit is used to summarize all holidays in a preset second time period based on a preset holiday summary strategy to obtain a holiday summary result, and calculate a daily weight index according to the holiday summary result and the second road rescue data.
[0078] Optionally, in one embodiment of the present application, the training module 103 includes: a prediction unit, a fourth acquisition unit and an optimization unit.
[0079] Among them, the prediction unit is used to train the daily rescue times prediction model according to the time series data derived variables, the location information derived variables and the weight index to obtain the training prediction score of the daily rescue times prediction model.
[0080] The fourth acquisition unit is used to obtain the predicted value and actual value of the daily rescue number prediction model when the training prediction score is not less than the preset score threshold, and train the safety redundancy threshold model based on the predicted value and the actual value to generate a safety redundancy threshold.
[0081] The optimization unit is used to optimize the distribution of road rescue resources through prediction values and safety redundancy thresholds.
[0082] According to the device for optimizing the distribution of road rescue resources in an offline training phase proposed in the embodiment of the present application, it includes a first acquisition module for collecting road rescue data of multiple vehicles corresponding to the target vehicle model; a generation module for generating time series data derived variables, position information derived variables and weight index corresponding to the road rescue data, wherein the weight index includes a weekly weight index and a daily weight index; a training module for training a preset daily rescue number prediction model and a safety redundancy threshold model based on the time series data derived variables, the position information derived variables and the weight index, so as to optimize the distribution of road rescue resources through the trained daily rescue number prediction model and the safety redundancy threshold model. The present application models and summarizes the road rescue situation from the perspective of time and space through road rescue historical data and driving conditions, so as to predict future road rescue conditions and provide reliable suggestions for the distribution of road rescue resources, so as to adjust and optimize the distribution of road rescue resources.
[0083] Figure 7 It is a block diagram of an apparatus for optimizing the distribution of road rescue resources in an online pre-stage according to an embodiment of the present application.
[0084] like Figure 7 As shown, the device 20 for comprehensively optimizing the distribution of road rescue resources in the online prediction stage includes: a second acquisition module 201 , an input module 202 and an adjustment module 203 .
[0085] The second acquisition module 201 is used to obtain target roadside assistance data corresponding to the target vehicle type.
[0086] The input module 202 is used to input the target road rescue data into a pre-trained daily rescue number prediction model to output the predicted rescue number in the target future time period, and obtain the safety redundancy threshold corresponding to the predicted rescue number based on the predicted rescue number and the pre-trained safety redundancy threshold model.
[0087] The adjustment module 203 is used to determine a road rescue resource distribution optimization suggestion based on the predicted rescue times and the safety redundancy threshold, so as to adjust the road rescue resource distribution through the road rescue resource distribution optimization suggestion.
[0088] Optionally, in one embodiment of the present application, the device 10 for comprehensively optimizing the distribution of road rescue resources in the online pre-stage of the present application also includes: a detection module, a judgment module, a stabilization module and a circulation module.
[0089] Among them, the detection module is used to detect the pre-trained daily rescue number prediction model and safety redundancy threshold model based on a preset detection cycle after adjusting the road rescue resource distribution through road rescue resource distribution optimization suggestions to obtain the corresponding KS indicator.
[0090] The judgment module is used to judge whether the KS indicator meets the preset stability conditions.
[0091] The stabilization module is used to optimize the distribution of road rescue resources by using the pre-trained daily rescue number prediction model and safety redundancy threshold model if the KS index meets the preset stability conditions.
[0092] The loop module is used to retrain the daily rescue number prediction model and the safety redundancy threshold model if the KS indicator does not meet the preset stability conditions.
[0093] It should be noted that the aforementioned explanation of the embodiment of the method for comprehensively optimizing the distribution of road rescue resources is also applicable to the device for comprehensively optimizing the distribution of road rescue resources in this embodiment, and will not be repeated here.
[0094] According to the device for optimizing the distribution of road rescue resources in the online prediction stage proposed in the embodiment of the present application, it includes a second acquisition module for obtaining target road rescue data corresponding to the target vehicle model; an input module for inputting the target road rescue data into a pre-trained daily rescue number prediction model to output the predicted number of rescues in the target future time period, and obtain the safety redundancy threshold corresponding to the predicted number of rescues based on the predicted number of rescues and the pre-trained safety redundancy threshold model; an adjustment module for determining optimization suggestions for the distribution of road rescue resources based on the predicted number of rescues and the safety redundancy threshold, so as to adjust the distribution of road rescue resources through the optimization suggestions for the distribution of road rescue resources. The present application models and summarizes the road rescue situation from the perspective of time and space through historical road rescue data and driving conditions, etc., so as to predict future road rescue conditions and provide reliable suggestions for the distribution of road rescue resources to adjust and optimize the distribution of road rescue resources.
[0095] Figure 8 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:
[0096] A memory 801 , a processor 802 , and a computer program stored in the memory 801 and executable on the processor 802 .
[0097] When the processor 802 executes the program, the method for comprehensively optimizing the distribution of road rescue resources provided in the above embodiment is implemented.
[0098] Furthermore, the electronic device further comprises:
[0099] The communication interface 803 is used for communication between the memory 801 and the processor 802 .
[0100] The memory 801 is used to store computer programs that can be executed on the processor 802 .
[0101] The memory 801 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.
[0102] If the memory 801, the processor 802 and the communication interface 803 are implemented independently, the communication interface 803, the memory 801 and the processor 802 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 8 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.
[0103] Optionally, in a specific implementation, if the memory 801, the processor 802 and the communication interface 803 are integrated on a chip, the memory 801, the processor 802 and the communication interface 803 can communicate with each other through an internal interface.
[0104] The processor 802 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.
[0105] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above method for comprehensively optimizing the distribution of road rescue resources.
[0106] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.
[0107] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.
[0108] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.
[0109] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.
[0110] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.
[0111] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.
[0112] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.
[0113] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.
Claims
1. A method for optimizing the distribution of road rescue resources, characterized in that: Applied in the offline training phase, it includes the following steps: Collect roadside assistance data of multiple vehicles corresponding to the target vehicle model; Generate time series data derived variables, location information derived variables and weight indexes corresponding to the road rescue data, wherein the weight indexes include weekly weight indexes and daily weight indexes; Based on the time series data derived variables, the location information derived variables and the weight index, a preset daily rescue times prediction model and a safety redundancy threshold model are trained to optimize the distribution of road rescue resources through the trained daily rescue times prediction model and the safety redundancy threshold model.
2. The method according to claim 1, characterized in that: The generating of the time series data derived variables, the location information derived variables and the weight index corresponding to the road rescue data includes: Acquire the first roadside assistance data of the plurality of vehicles on each non-holiday within a preset first time period; Acquire second roadside assistance data of the plurality of vehicles on each holiday within a preset second time period; generating the time series data derived variable and the position information derived variable according to the first roadside assistance data and the second roadside assistance data; Based on a preset weekly summary strategy, all non-holidays in the preset first time period are summarized to obtain a non-holiday summary result, and the weekly weight index is calculated according to the weekly summary result and the first roadside assistance data; Based on a preset holiday summary strategy, all holidays in the preset second time period are summarized to obtain a holiday summary result, and the daily weight index is calculated based on the holiday summary result and the second roadside assistance data.
3. The method according to claim 1, characterized in that The method of training a preset daily rescue number prediction model and a safety redundancy threshold model based on the time series data derived variable, the location information derived variable and the weight index, so as to optimize the distribution of road rescue resources through the trained daily rescue number prediction model and the safety redundancy threshold model, includes: Training the daily rescue times prediction model according to the time series data derived variables, the location information derived variables and the weight index to obtain a training prediction score of the daily rescue times prediction model; When the training prediction score is not less than the preset score threshold, obtaining the predicted value and the actual value of the daily rescue number prediction model, and training the safety redundancy threshold model based on the predicted value and the actual value to generate a safety redundancy threshold; The road rescue resource distribution is optimized by using the predicted value and the safety redundancy threshold.
4. A method for optimizing the distribution of road rescue resources, characterized in that: Applied to the online prediction stage, it includes the following steps: Obtain target roadside assistance data corresponding to the target vehicle type; Inputting the target road rescue data into a pre-trained daily rescue number prediction model to output the predicted rescue number in a target future time period, and obtaining a safety redundancy threshold value corresponding to the predicted rescue number based on the predicted rescue number and a pre-trained safety redundancy threshold model; Based on the predicted number of rescue times and the safety redundancy threshold, a road rescue resource distribution optimization suggestion is determined, so as to adjust the road rescue resource distribution through the road rescue resource distribution optimization suggestion.
5. The method according to claim 4, characterized in that After adjusting the road rescue resource distribution by using the road rescue resource distribution optimization suggestion, the method further includes: Based on a preset detection cycle, the pre-trained daily rescue number prediction model and safety redundancy threshold model are tested to obtain corresponding KS indicators; Determine whether the KS indicator meets the preset stability condition; If the KS index satisfies the preset stability condition, the pre-trained daily rescue number prediction model and safety redundancy threshold model are continued to be used to optimize the road rescue resource distribution; If the KS indicator does not meet the preset stability condition, the daily rescue number prediction model and the safety redundancy threshold model are retrained.
6. A device for optimizing the distribution of road rescue resources, characterized in that: Applied in the offline training phase, including: A first collection module is used to collect roadside assistance data of multiple vehicles corresponding to a target vehicle type; A generating module, used to generate time series data derived variables, location information derived variables and weight indexes corresponding to the road rescue data, wherein the weight indexes include weekly weight indexes and daily weight indexes; A training module is used to train a preset daily rescue number prediction model and a safety redundancy threshold model based on the time series data derived variables, the location information derived variables and the weight index, so as to optimize the distribution of road rescue resources through the trained daily rescue number prediction model and the safety redundancy threshold model.
7. The device according to claim 6, characterized in that The generation module comprises: A first acquisition unit is used to acquire first roadside assistance data of the plurality of vehicles on each non-holiday within a preset first time period; A second acquisition unit is used to acquire second roadside assistance data of the plurality of vehicles on each holiday within a preset second time period; a third acquisition unit, configured to generate the time series data derived variable and the position information derived variable according to the first road rescue data and the second road rescue data; A first summarizing unit is configured to summarize all non-holidays in the preset first time period based on a preset weekly summary strategy to obtain a non-holiday summary result, and calculate the weekly weight index according to the weekly summary result and the first roadside assistance data; The second summarizing unit is used to summarize all holidays in the preset second time period based on a preset holiday summary strategy to obtain a holiday summary result, and calculate the daily weight index according to the holiday summary result and the second road rescue data.
8. A device for optimizing the distribution of road rescue resources, characterized in that: Applied to the online prediction stage, including: The second acquisition module is used to obtain target roadside assistance data corresponding to the target vehicle type; An input module, used to input the target road rescue data into a pre-trained daily rescue number prediction model to output the predicted rescue number in a target future time period, and obtain a safety redundancy threshold value corresponding to the predicted rescue number based on the predicted rescue number and a pre-trained safety redundancy threshold model; The adjustment module is used to determine a road rescue resource distribution optimization suggestion based on the predicted number of rescue times and the safety redundancy threshold, so as to adjust the road rescue resource distribution through the road rescue resource distribution optimization suggestion.
9. A vehicle, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for optimizing the distribution of road rescue resources as described in any one of claims 1-3 or 4-5.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement a method for comprehensively optimizing the distribution of road rescue resources as described in any one of claims 1-3 or 4-5.
