Dynamic resource allocation method based on cloud computing

Abstract

The invention discloses a resource dynamic allocation method based on cloud computing. The method comprises the following steps: step 1, realizing sub-region division and labeling through grid and map data; 2, realizing user order prediction by establishing a regional taxi-taking order prediction model; 3, predicting a user order to minimize the empty driving time of a driver; 4, order priority ranking is carried out, user feedback is carried out, a taxi-taking order prediction and delivery optimization module is used for establishing a taxi-taking order prediction model of a region level, and the utilization rate of the driver vehicle is improved through an optimized delivery algorithm; the intelligent distribution and real-time optimization module is used for minimizing the empty driving time of a driver through an intelligent order dispatching algorithm, and realizing intelligent distribution and real-time optimization of orders in combination with a real-time order demand map, traffic information and a dynamic adjustment strategy; the method has the characteristics of improving the vehicle utilization rate and the taxi taking experience of the user.

Description

Technical Field

[0001] The present invention relates to the technical field of resource allocation, and in particular to a method for dynamic resource allocation based on cloud computing. Background Art

[0002] In the existing taxi service system, especially in busy areas and during peak hours, users often face extremely inconvenient long waits and frequent queues. The current traditional system is unable to effectively cope with the surge in traffic demand during peak hours, resulting in a sharp decline in the overall user experience. At the same time, it also causes drivers to face a considerable amount of idle waiting time, which greatly reduces the overall vehicle utilization rate.

[0003] Under this situation, traditional taxi service systems have failed to fully utilize cloud computing and advanced technologies to intelligently allocate and optimize orders. Problems include the system's failure to accurately predict peak order demand, resulting in uneven vehicle allocation and difficulty in meeting user needs. At the same time, the common queuing phenomenon in busy areas also occurs frequently. In addition, while drivers are waiting for orders, there is a lot of idle time, which not only greatly reduces the time utilization of vehicles, but also creates unnecessary burdens on the environment. This phenomenon is particularly significant in busy urban traffic, making traditional taxi service systems appear inefficient and inefficient in high-load and high-demand environments. Therefore, it is necessary to design a cloud computing-based resource dynamic allocation method to improve vehicle utilization and user taxi experience. Summary of the invention

[0004] The purpose of the present invention is to provide a cloud computing-based dynamic resource allocation method to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a resource dynamic allocation system based on cloud computing, the operation method of the system comprises the following steps:

[0006] Step 1: Divide and label sub-areas using grid and map data;

[0007] Step 2: Realize user order prediction by establishing a regional taxi order prediction model;

[0008] Step 3: Minimize the driver’s idle time by predicting user orders;

[0009] Step 4: Rank order priorities and implement user feedback.

[0010] According to the above technical solution, the step of implementing sub-area division and labeling through grid and map data includes:

[0011] The system first collects satellite images, map data and real-time traffic information of the current area, and then pre-processes the collected data and divides it into sub-areas. The specific method is as follows: first select a square grid shape, then project the map data into the coordinate system, then divide the map into square grids, and finally associate each grid with its geographic coordinates so that each grid has a unique identifier. After the grid division is completed, based on the historical order quantity of each sub-area, each sub-area is marked as high demand and low demand. By counting the historical order quantity of each sub-area, the system calculates the average order quantity and uses this as the threshold. If the order quantity of a sub-area is higher than the average historical order quantity, it is marked as a high demand area; otherwise, it is marked as a low demand area.

[0012] According to the above technical solution, the step of realizing user order prediction by establishing a regional taxi order prediction model includes:

[0013] Extract key features related to taxi demand;

[0014] Deploy the trained model to the actual taxi-hailing system to predict the taxi demand at each location and time period.

[0015] According to the above technical solution, the step of extracting key features related to taxi demand includes:

[0016] A dedicated storage space is established on the cloud computing platform to store historical taxi order data obtained from various available data sources, including time, location, and order completion time information. Subsequently, feature engineering is performed on the collected data to mainly extract the following features. First, time features are extracted using timestamps, mainly extracting hours and dates. Second, order duration is calculated by using timestamp data to calculate the time difference between order creation time and completion time to obtain order duration information. Time series analysis methods are used to perform sliding time window statistics on the number of historical orders on the cloud computing platform, which includes using a statistical library to count the number of orders in each time window. Finally, a machine learning or deep learning model is used to train an event recognition model to identify special events in historical order data, such as National Day holidays.

[0017] According to the above technical solution, the step of deploying the trained model to the actual taxi-hailing system to realize the prediction of taxi demand at each location and time period includes:

[0018] The dataset is divided into a training set and a test set. The dataset is the data features extracted in the above steps, of which 80% of the data is used for model training and 20% is used to evaluate model performance. The GBM model is selected and key parameters are configured, including 100 trees, a learning rate of 0.1, and a maximum depth of 3 for each tree. The selection of these parameters is based on the understanding of the problem and preliminary experimental results. Next, the GBM model is trained using the training set. The model gradually learns the complex relationships in the historical order data to more accurately predict future taxi demand. After the training is completed, the model is applied to the test set. The performance of the model on the test data is measured by evaluation indicators such as the root mean square error. By analyzing the output of the model, it is possible to understand which features are most important for predicting taxi demand. Finally, the model is verified using the validation set to ensure its generalization ability on new data. Finally, when the model performs well in the test phase, it is deployed to the actual taxi system. In real-time applications, the model can provide predictions of taxi demand for each location and time period, thereby helping the system to intelligently allocate driver resources and improve service efficiency and user experience.

[0019] According to the above technical solution, the step of minimizing the driver's idle time by predicting the user's orders includes:

[0020] Set minimizing driver idle time as the optimization goal;

[0021] Integrate real-time traffic information to provide drivers with the best route to avoid congested roads.

[0022] According to the above technical solution, the step of setting minimization of driver idle time as the optimization goal includes:

[0023] The optimization goal of the intelligent dispatching algorithm can be expressed by the following mathematical formula, which takes into account minimizing the driver's idle time: Idle time i , the objective function includes the idle time of all drivers, n is the number of drivers, and the idle time i Represents the idle time of the i-th driver. When dispatching orders, the system monitors order demand and driver location in real time to ensure that the driver can quickly receive the next order after completing the order. This involves real-time decision-making of the intelligent dispatching algorithm to guide the driver to the area most likely to have new orders. The decision of the algorithm is based on the real-time nature of the order, the geographical location and the output of the prediction model. The system uses a real-time order demand map at the system level to dynamically monitor the order situation in each area to predict places with high order density, thereby reducing the driver's idle waiting time.

[0024] According to the above technical solution, the step of integrating real-time traffic information to provide drivers with the best route to avoid congested roads includes:

[0025] At the same time, the system efficiently processes and utilizes traffic flow data and real-time map information through cloud computing resources, and combines it with the intelligent navigation system to provide drivers with the optimal route to avoid congested roads. In the formula, it is expressed as: Driving time j , where m is the number of road segments on the path and the travel time j represents the travel time of the jth road section. Finally, the system monitors the order status and driver location in real time, and dynamically adjusts the delivery strategy to adapt to different traffic and order demand conditions. During peak hours or special events, the delivery algorithm is adjusted immediately and the driver allocation strategy is re-evaluated to cope with traffic congestion or order surges.

[0026] According to the above technical solution, the steps of ranking order priorities and implementing user feedback include:

[0027] Through the prediction model trained in step 2, the system detects that a sub-region is about to enter a peak period, and the system immediately begins to guide orders for low-demand areas. Specifically, the implementation method of order guidance includes ranking the priorities of orders. The system will assign priorities to each order based on the distance from the high-demand area. Orders closer to the high-demand area will have higher priorities. In addition, the system also refers to the priority of the queuing time to avoid long waiting times, that is: priority = α×distance priority+(1-α)×time priority, where α is the weight for weighing distance and time, which can be adjusted according to actual conditions. Combined with this priority information, the system performs intelligent dispatching to ensure that orders in low-demand areas are given priority during the order guidance process. This requires appropriate adjustments to the intelligent dispatching algorithm to make it more flexible during order guidance.

[0028] According to the above technical solution, the system includes:

[0029] The regional division and labeling module is used to provide a reliable geographical basis for the subsequent dynamic allocation of cloud computing-based resources, ensuring that the system can intelligently identify and respond to the needs of different sub-regions;

[0030] The taxi order prediction and delivery optimization module is used to establish a regional-level taxi order prediction model and improve the utilization rate of drivers and vehicles by optimizing the delivery algorithm;

[0031] The intelligent allocation and real-time optimization module is used to minimize the driver's idle time through an intelligent dispatching algorithm, and to achieve intelligent allocation and real-time optimization of orders by combining real-time order demand maps, traffic information and dynamic adjustment strategies.

[0032] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: the present invention realizes accurate regional division and high and low demand marking by collecting map data, dividing sub-regions, and combining cloud computing resources. By establishing a regional-level order prediction model and optimizing the delivery algorithm, the cloud computing power is fully utilized to improve the utilization rate of drivers and vehicles. The system also introduces an order guidance mechanism and dynamic vehicle allocation, detects high-demand sub-regions according to the prediction model and flexibly allocates vehicle resources. Finally, through user feedback data and cloud analysis, the prediction model and delivery algorithm are continuously optimized to improve the system's accurate response to user needs and optimize user experience. This entire system architecture and method, based on the cloud computing platform, provides an efficient, intelligent and personalized solution for taxi services. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0034] Figure 1 A flowchart of a method for dynamic resource allocation based on cloud computing provided in Embodiment 1 of the present invention;

[0035] Figure 2 A schematic diagram of the module composition of a cloud computing-based dynamic resource allocation system provided in Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] Embodiment 1: Figure 1 This is a flow chart of a method for dynamic resource allocation based on cloud computing provided in Example 1 of the present invention. This embodiment can be applied to the scenario of online car-hailing order allocation. This method can be executed by a dynamic resource allocation system based on cloud computing provided in this embodiment. Figure 1 As shown, the method specifically comprises the following steps:

[0038] Step 1: Divide and label sub-areas using grid and map data;

[0039] In the embodiment of the present invention, a regional map is collected and divided into sub-regions, and then the sub-regions are divided into high-demand and low-demand regions, providing a reliable geographical basis for subsequent dynamic resource allocation based on cloud computing;

[0040] Exemplarily, the system first collects satellite images, map data and real-time traffic information of the current area, and then pre-processes the collected data and divides it into sub-areas. The specific method is as follows: first, a square grid shape is selected, and then the map data is projected into a coordinate system. Then, the map is divided into square grids. Finally, each grid is associated with its geographic coordinates so that each grid has a unique identifier. After the grid division is completed, each sub-area is marked as high demand or low demand based on the historical order quantity of each sub-area. By counting the historical order quantity of each sub-area, the system calculates the average order quantity and uses it as a threshold. If the order quantity of a sub-area is higher than the average historical order quantity, it is marked as a high demand area; otherwise, it is marked as a low demand area. This step uses the average order quantity as a threshold to enable the system to divide geographical areas into high demand and low demand, providing an accurate basis for subsequent dynamic allocation of resources based on cloud computing.

[0041] Step 2: Realize user order prediction by establishing a regional taxi order prediction model;

[0042] In the embodiment of the present invention, in order to realize the prediction of taxi orders, a regional-level taxi order prediction model is established to provide a basis for subsequent order allocation;

[0043] Exemplarily, a dedicated storage space is established on the cloud computing platform to store historical taxi order data obtained from various available data sources, including important information such as time, location, and order completion time. Subsequently, feature engineering is performed on the collected data to mainly extract the following features: first, time feature extraction is performed using timestamps, mainly extracting hours and dates; second, order duration calculation is performed, using timestamp data to calculate the time difference between order creation time and completion time to obtain order duration information; and using time series analysis methods, a sliding time window of historical order quantity is counted on the cloud computing platform, which includes using a statistical library to count the number of orders in each time window; finally, a machine learning or deep learning model is used to train an event recognition model to identify special events in historical order data, such as National Day holidays;

[0044] Exemplarily, the data set is divided into a training set and a test set. The data set is the data features extracted in the above steps, of which 80% of the data is used for model training and 20% is used to evaluate model performance. The GBM model is selected and key parameters are configured, including 100 trees, a learning rate of 0.1, and a maximum depth of 3 for each tree. The selection of these parameters is based on the understanding of the problem and preliminary experimental results. Next, the GBM model is trained using the training set. The model gradually learns the complex relationships in the historical order data to more accurately predict future taxi demand. After the training is completed, the model is applied to the test set. The performance of the model on the test data is measured by evaluation indicators such as the root mean square error. Moreover, by analyzing the output of the model, it can be understood which features are most important for predicting taxi demand. Finally, the model is verified using the validation set to ensure its generalization ability on new data. Finally, when the model performs well in the test phase, it is deployed to the actual taxi system. In real-time applications, the model can provide predictions of taxi demand for each location and time period, thereby helping the system to intelligently allocate driver resources and improve service efficiency and user experience.

[0045] Step 3: Minimize the driver’s idle time by predicting user orders;

[0046] In the embodiment of the present invention, the order quantity is predicted based on step 2, and the driver's idle driving time is minimized through an intelligent dispatch optimization algorithm to improve the overall vehicle utilization rate;

[0047] Exemplarily, the optimization objective of the intelligent dispatching algorithm can be expressed by the following mathematical formula, which takes into account minimizing the driver's idle time: Idle time i , the objective function includes the idle time of all drivers, n is the number of drivers, and the idle time i represents the idle time of the i-th driver. When dispatching orders, the system monitors order demand and driver location in real time to ensure that the driver can quickly receive the next order after completing the order. This involves the real-time decision-making of the intelligent dispatching algorithm to guide the driver to the area most likely to have new orders. The decision of the algorithm is based on the real-time nature of the order, the geographical location and the output of the prediction model. The system uses a real-time order demand map at the system level to dynamically monitor the order situation in each area to predict places with high order density, thereby reducing the idle waiting time of the driver. This process requires the use of a real-time order demand map combined with the output of the prediction model to achieve dynamic adjustment of each area;

[0048] Exemplarily, at the same time, the system efficiently processes and utilizes traffic flow data and real-time map information through cloud computing resources, combined with the intelligent navigation system, to provide drivers with the best path to avoid congested sections, which is expressed in the formula as: Driving timej , where m is the number of road segments on the path and the travel time j represents the travel time of the jth road section. Finally, the system monitors the order status and driver location in real time, and dynamically adjusts the delivery strategy to adapt to different traffic and order demand conditions. During peak hours or special events, the delivery algorithm is adjusted immediately and the driver allocation strategy is re-evaluated to cope with traffic congestion or order surges.

[0049] Step 4: Rank order priorities and implement user feedback.

[0050] In the embodiment of the present invention, once the system detects that a high-demand sub-area is about to usher in an order peak period, the system will activate the order guidance mechanism to effectively allocate vehicle resources;

[0051] Exemplarily, first, the system detects that a sub-region is about to enter a peak period through the prediction model trained in step 2, and the system immediately starts to guide orders for low-demand areas. Specifically, the implementation method of order guidance includes ranking the priority of orders. The system will assign a priority to each order according to the distance from the high-demand area. Orders closer to the high-demand area will have a higher priority. In addition, the system also refers to the priority of the queuing time to avoid long waiting times, that is, priority = α×distance priority+(1-α)×time priority, where α is the weight for weighing distance and time, which can be adjusted according to actual conditions. Combined with these priority information, the system performs intelligent dispatching to ensure that orders in low-demand areas are given priority during the order guidance process. This requires appropriate adjustments to the intelligent dispatching algorithm to make it more flexible during the order guidance period;

[0052] Exemplarily, a user feedback module is introduced to allow passengers to provide feedback on waiting time, vehicle arrival time, driver service quality, etc. The collected feedback data is uploaded to the cloud. Then, cloud computing resources are used to perform real-time analysis of the user feedback data, including the use of machine learning algorithms to identify key trends and patterns, such as which factors are associated with user dissatisfaction experiences. Through these analysis results, the prediction model and dispatch algorithm can be continuously improved to more accurately meet user needs. In addition, through cloud computing, system performance is monitored in real time, with special attention paid to key user experience indicators such as waiting time and vehicle arrival time. Once performance degradation or poor user feedback is detected, the system will immediately take corresponding measures. Based on the results of cloud analysis and performance monitoring, system parameters and algorithms can be adjusted in real time to optimize dispatch strategies and improve system performance. This ensures that the system can continuously adapt to the ever-changing operating environment to ensure that users get the best service experience.

[0053] Embodiment 2: Embodiment 2 of the present invention provides a resource dynamic allocation system based on cloud computing. Figure 2 The schematic diagram of the module composition of the resource dynamic allocation system based on cloud computing provided in the second embodiment of the present invention is as follows: Figure 2 As shown, the system includes:

[0054] The regional division and labeling module is used to provide a reliable geographical basis for the subsequent dynamic allocation of cloud computing-based resources, ensuring that the system can intelligently identify and respond to the needs of different sub-regions;

[0055] The taxi order prediction and delivery optimization module is used to establish a regional-level taxi order prediction model and improve the utilization rate of drivers and vehicles by optimizing the delivery algorithm;

[0056] Intelligent allocation and real-time optimization module, which is used to minimize the idle time of drivers through intelligent dispatching algorithms, and realize intelligent allocation and real-time optimization of orders by combining real-time order demand maps, traffic information and dynamic adjustment strategies;

[0057] In some embodiments of the present invention, the region division and labeling module includes:

[0058] The data collection and preprocessing module is used to collect satellite images, map data and real-time traffic information, and use the cloud computing platform to perform data preprocessing, including map projection and grid division, to ensure that each sub-area has a unique identifier;

[0059] The sub-region demand labeling module is used to label high and low demand for each sub-region based on the historical order quantity through the cloud computing platform;

[0060] The resource dynamic allocation module is used to provide a reliable geographical basis for the subsequent dynamic allocation of cloud computing resources;

[0061] In some embodiments of the present invention, the taxi order prediction and delivery optimization module includes:

[0062] The data storage and feature extraction module is used to establish storage space on the cloud computing platform to store historical order data, perform feature engineering, and extract time and event features, including timestamps, order duration, and identification of special events;

[0063] Model training and optimization module, which uses cloud computing power to train the taxi order prediction model, selects appropriate machine learning models, such as GBM, and ensures model performance through cross-validation and validation sets;

[0064] The system real-time application module is used to deploy the trained model to the actual system to achieve real-time prediction of taxi demand at each location and time period, helping the system to dispatch orders intelligently and improve service efficiency;

[0065] In some embodiments of the present invention, the intelligent allocation and real-time optimization module includes:

[0066] The dispatch algorithm and route planning module is used to minimize the driver's idle time through an intelligent dispatch algorithm, taking into account the real-time nature of orders, geographic location, and prediction model output, and provide drivers with the best route to avoid congested sections;

[0067] Traffic information utilization and dynamic adjustment strategy module, which is used to integrate real-time traffic information to prevent drivers from entering congested areas, efficiently process and utilize this information through cloud computing resources, monitor order status and driver location in real time, and dynamically adjust delivery strategies to adapt to different traffic and order demand conditions;

[0068] The real-time system monitoring and adjustment module is used to use cloud computing resources to monitor system performance in real time, including key user experience indicators. Once performance degradation or poor user feedback is detected, the system will immediately adjust parameters and algorithms to optimize the delivery strategy and ensure efficient operation of the system.

[0069] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0070] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for dynamic resource allocation based on cloud computing, characterized in that: The method comprises the following steps: Step 1: Divide and label sub-areas using grid and map data; Step 2: Realize user order prediction by establishing a regional taxi order prediction model; Step 3: Minimize the driver’s idle time by predicting user orders; Step 4: Rank order priorities and implement user feedback.

2. The method for dynamic resource allocation based on cloud computing according to claim 1, characterized in that: The step of implementing sub-area division and labeling through grid and map data includes: The system first collects satellite images, map data and real-time traffic information of the current area, and then pre-processes the collected data and divides it into sub-areas. The specific method is as follows: first select a square grid shape, then project the map data into the coordinate system, then divide the map into square grids, and finally associate each grid with its geographic coordinates so that each grid has a unique identifier. After the grid division is completed, based on the historical order quantity of each sub-area, each sub-area is marked as high demand and low demand. By counting the historical order quantity of each sub-area, the system calculates the average order quantity and uses this as the threshold. If the order quantity of a sub-area is higher than the average historical order quantity, it is marked as a high demand area; otherwise, it is marked as a low demand area.

3. The method for dynamic resource allocation based on cloud computing according to claim 1, characterized in that: The step of realizing user order prediction by establishing a regional taxi order prediction model comprises: Extract key features related to taxi demand; Deploy the trained model to the actual taxi-hailing system to predict the taxi demand at each location and time period.

4. The method for dynamic resource allocation based on cloud computing according to claim 3, characterized in that: The step of extracting key features related to taxi demand includes: A dedicated storage space is established on the cloud computing platform to store historical taxi order data obtained from various available data sources, including time, location, and order completion time information. Subsequently, feature engineering is performed on the collected data to mainly extract the following features. First, time features are extracted using timestamps, mainly extracting hours and dates. Second, order duration is calculated by using timestamp data to calculate the time difference between order creation time and completion time to obtain order duration information. Time series analysis methods are used to perform sliding time window statistics on the number of historical orders on the cloud computing platform, which includes using a statistical library to count the number of orders in each time window. Finally, a machine learning or deep learning model is used to train an event recognition model to identify special events in historical order data, such as National Day holidays.

5. The method for dynamic resource allocation based on cloud computing according to claim 3, characterized in that: The step of deploying the trained model to the actual taxi-hailing system to predict the taxi demand at each location and time period includes: The dataset is divided into a training set and a test set. The dataset is the data features extracted in the above steps, of which 80% of the data is used for model training and 20% is used to evaluate model performance. The GBM model is selected and key parameters are configured, including 100 trees, a learning rate of 0.1, and a maximum depth of 3 for each tree. The selection of these parameters is based on the understanding of the problem and preliminary experimental results. Next, the GBM model is trained using the training set. The model gradually learns the complex relationships in the historical order data to more accurately predict future taxi demand. After the training is completed, the model is applied to the test set. The performance of the model on the test data is measured by evaluation indicators such as the root mean square error. By analyzing the output of the model, it is possible to understand which features are most important for predicting taxi demand. Finally, the model is verified using the validation set to ensure its generalization ability on new data. Finally, when the model performs well in the test phase, it is deployed to the actual taxi system. In real-time applications, the model can provide predictions of taxi demand for each location and time period, thereby helping the system to intelligently allocate driver resources and improve service efficiency and user experience.

6. The method for dynamic resource allocation based on cloud computing according to claim 1, characterized in that: The step of minimizing the driver's idle time by predicting the user's orders includes: Set minimizing driver idle time as the optimization goal; Integrate real-time traffic information to provide drivers with the best route to avoid congested roads.

7. The method for dynamic resource allocation based on cloud computing according to claim 6, characterized in that: The step of setting minimization of driver idle time as the optimization goal includes: The optimization goal of the intelligent dispatching algorithm can be expressed by the following mathematical formula, which takes into account minimizing the driver's idle time: Idle time i , the objective function includes the idle time of all drivers, n is the number of drivers, and the idle time i Represents the idle time of the i-th driver. When dispatching orders, the system monitors order demand and driver location in real time to ensure that the driver can quickly receive the next order after completing the order. This involves real-time decision-making of the intelligent dispatching algorithm to guide the driver to the area most likely to have new orders. The decision of the algorithm is based on the real-time nature of the order, the geographical location and the output of the prediction model. The system uses a real-time order demand map at the system level to dynamically monitor the order situation in each area to predict places with high order density, thereby reducing the driver's idle waiting time.

8. The method for dynamic resource allocation based on cloud computing according to claim 6, characterized in that: The step of integrating real-time traffic information to provide the driver with an optimal route to avoid congested roads includes: At the same time, the system efficiently processes and utilizes traffic flow data and real-time map information through cloud computing resources, and combines it with the intelligent navigation system to provide drivers with the optimal route to avoid congested roads. In the formula, it is expressed as: Driving time j , where m is the number of road segments on the path and the travel time j represents the travel time of the jth road section. Finally, the system monitors the order status and driver location in real time, and dynamically adjusts the delivery strategy to adapt to different traffic and order demand conditions. During peak hours or special events, the delivery algorithm is adjusted immediately and the driver allocation strategy is re-evaluated to cope with traffic congestion or order surges.

9. The method for dynamic resource allocation based on cloud computing according to claim 1, characterized in that: The steps of ranking order priorities and implementing user feedback include: The system detects that a sub-region is about to enter a peak period through the prediction model trained in step 2. The system then starts to guide orders for low-demand areas. Specifically, the implementation of order guidance includes ranking the priority of orders. The system will assign a priority to each order based on the distance from the high-demand area. Orders closer to the high-demand area will have a higher priority. In addition, the system also refers to the priority of the queuing time to avoid long waiting times, that is, priority = α× Distance priority + (1-α) × time priority, where α is the weight for weighing distance and time, which can be adjusted according to actual conditions. Combined with these priority information, the system performs intelligent dispatching to ensure that orders from low-demand areas are given priority during the order guidance process. This requires moderate adjustments to the intelligent dispatching algorithm to make it more flexible during order guidance.

10. The cloud computing-based dynamic resource allocation method according to claim 1 is applied to a cloud computing-based dynamic resource allocation system, characterized in that: The system comprises: The regional division and labeling module is used to provide a reliable geographical basis for the subsequent dynamic allocation of cloud computing-based resources, ensuring that the system can intelligently identify and respond to the needs of different sub-regions; The taxi order prediction and delivery optimization module is used to establish a regional-level taxi order prediction model and improve the utilization rate of drivers and vehicles by optimizing the delivery algorithm; The intelligent allocation and real-time optimization module is used to minimize the driver's idle time through an intelligent dispatching algorithm, and to achieve intelligent allocation and real-time optimization of orders by combining real-time order demand maps, traffic information and dynamic adjustment strategies.

Images