Vehicle-mounted edge computing environment-oriented time delay optimization task unloading and resource allocation parallel method and system

Abstract

The invention discloses a time delay optimization task unloading and resource allocation parallel method and system oriented to a vehicle-mounted edge computing environment. The method comprises the steps that S1, a model for cooperative computing of a target vehicle and an idle vehicle under RSU coverage is established; s2, configuring motion models of the target vehicle and the idle vehicle, calculation and unloading options of the target vehicle, a task allocation proportion and a communication range; s3, giving a transmission rate formula to obtain task transmission delay; s4, calculating task time delay; s5, by taking minimization of parallel system time delay as a target, establishing an optimization model of joint task allocation and time delay minimization according to a task allocation proportion constraint and a time delay constraint of vehicle task calculation unloading; and S6, solving the task allocation proportion by enabling the time delay of local processing of the target vehicle, the time delay of processing of the mobile edge computing server equipped by the road side unit and the time delay of processing of the idle vehicle to be equal. Therefore, the minimum time delay of the target vehicle for solving the task is obtained.

Description

Technical Field

[0001] The present invention belongs to the field of wireless communication technology, and specifically relates to a parallel method and system for delay optimization task offloading and resource allocation for an on-board edge computing environment. Background Art

[0002] With the rapid development of wireless communication technology, demanding services and applications will further test cellular and wireless networks. Mobile edge computing, as a promising technology for providing computing services in mobile edge networks, has been proposed to solve the above problems. Mobile edge computing deploys computing and storage resources at the edge of mobile networks to provide IT service environment and cloud computing capabilities for mobile networks, thereby providing users with ultra-low latency and high-bandwidth network service solutions, ultimately minimizing system latency or energy consumption, etc.

[0003] Research on mobile edge computing offloading resource allocation algorithms has made great progress. However, most of these studies consider the task as a whole, do not split it, and only select a single device for offloading calculations. In fact, the computing power of each computing device is limited. Assigning tasks to a single device for calculation will result in longer delays. After reasonably considering issues such as task segmentation, scheduling, and transmission, tasks are reasonably allocated to multiple computing devices, and each part is processed in parallel to achieve higher efficiency. In addition, the edge computing nodes of these studies are usually fixed, and the dynamic nature of edge computing nodes in real life is not considered. Based on this, the present invention proposes a parallel method and system for delay-optimized task offloading and resource allocation for an on-board edge computing environment. Summary of the invention

[0004] In view of the above-mentioned defects in the prior art, based on the mobile vehicle edge computing network, the present invention discloses a parallel method and system for delay-optimized task offloading and resource allocation for vehicle-mounted edge computing environments. Aiming at the application scenario of mobile vehicles, the present invention takes minimizing the delay of the parallel system as the goal, establishes an optimization model for joint task allocation and delay minimization according to the task allocation ratio constraints and the delay constraints of vehicle task calculation offloading, and uses a task allocation strategy that makes the task processing time on the local and offloading devices equal to achieve the optimal allocation of task ratios.

[0005] In order to achieve the purpose of the present invention, the present invention adopts the following technical scheme:

[0006] A parallel method for delay optimization task offloading and resource allocation for an on-vehicle edge computing environment includes the following steps:

[0007] S1. Model establishment: Establish a model for collaborative calculation between target vehicles and idle vehicles under the coverage of a road side unit (RSU);

[0008] S2, initial system configuration: motion model of target vehicle and idle vehicles, calculation and unloading options of target vehicle, task allocation ratio and communication range;

[0009] S3. Analyze the communication model: give the transmission rate formula and obtain the task transmission delay;

[0010] S4, Analyze the calculation model: calculate the task delay;

[0011] S5. Establishing an optimization model: with the goal of minimizing the delay of the parallel system, an optimization model for joint task allocation and delay minimization is established according to the task allocation ratio constraint and the delay constraint of vehicle task calculation offloading;

[0012] S6. Solve the optimization model: Solve the task allocation ratio by making the delay of local processing of the target vehicle, the delay of processing by the mobile edge computing server (MEC server) equipped with the roadside unit, and the delay of processing by the idle vehicle equal.

[0013] Furthermore, in step S1, the model established is specifically:

[0014] Consider a one-way straight road with roadside units deployed along one side of the road. In addition to the target vehicle, there is also an idle friendly assistance vehicle in the system. They are all traveling in the same direction under the coverage of RSU.

[0015] Furthermore, in step S2, the initial system configuration is:

[0016] The vehicle travels at a constant speed, the communication coverage of the RSU is considered as a cell, and the RSU is equipped with a fixed mobile edge computing server as an edge computing node. Consider that the target vehicle starts from the left end of the RSU coverage area and drives out of the rightmost end of the coverage area. The radius of the RSU is R. At the junction of the road and the RSU coverage area, the angle between the road and the RSU radius is θ, so the vehicle travels the distance under the RSU as:

[0017] K=2Rcosθ

[0018] The task generated on the target vehicle is represented by <D, F>, where D represents the input data size and F represents the CPU cycles required to complete the task. The target vehicle can choose to perform its task locally or offload it to the MEC server of the RSU and execute it in parallel. The proportion of tasks assigned to each device is recorded as A = {k0, k1, k2}. k0, k1, k2 represent the proportion of tasks calculated locally and offloaded to the RSU (MEC server) and unloaded to idle vehicles, respectively. The proportion of these three parts is weighted as 1, which can be expressed as:

[0019] k0+k1+k2=1

[0020] According to the allocation ratio, the task size allocated to local computing and the number of CPU cycles required to complete this task can be expressed as:

[0021]

[0022] The task size assigned to the MEC server on the RSU and the number of CPU cycles required to complete this task are expressed as:

[0023]

[0024] The size of the task assigned to the idle vehicle and the number of CPU cycles required to complete this task are expressed as:

[0025]

[0026] Furthermore, in step S3, the communication model is specifically:

[0027] Orthogonal frequency division multiplexing technology is used. Therefore, the data transmission rate from the target vehicle to the roadside unit (equipped with a MEC server on the RSU) can be expressed as:

[0028]

[0029] The data transmission rate between the target vehicle and the idle vehicle can be expressed as:

[0030]

[0031] Among them, B M , B V2V They represent the channel bandwidth of the uplink between the target vehicle and the RSU (MEC server) and the idle vehicle, respectively, P represents the transmission power of the target vehicle device data, and d TM d V2V They represent the distances between the target vehicle and the RSU (MEC server) and between the target vehicle and the idle vehicle, ζ is the path loss index, and h M and h V2V Respectively represent the channel gain between the target vehicle and the RSU (MEC server) and the idle vehicle, and N0 is the additive white noise power. In order to simplify the calculation, R is used to represent d TM , expressed as:

[0032] d TM =R

[0033] The transmission delay of the partial tasks generated by the target vehicle uploaded to the RSU (MEC server) is expressed as:

[0034]

[0035] The transmission delay of uploading part of the tasks generated by the target vehicle to the idle vehicle is expressed as:

[0036]

[0037] Furthermore, in step S4, the calculation model is specifically:

[0038] The computing tasks <D, F> generated on the target vehicle can be simultaneously assigned to the local vehicle, the MEC server on the RSU, and the idle vehicle SV for parallel processing. The corresponding local computing latency is:

[0039]

[0040] The task calculation delay transmitted to the RSU (MEC server) is:

[0041]

[0042] The task calculation delay transmitted to the idle vehicle part is:

[0043]

[0044] Among them, f L 、f M 、f S They represent the computing capabilities of the target vehicle, the MEC server equipped by the RSU, and the idle vehicles respectively.

[0045] Furthermore, in step S5, the optimization problem is:

[0046] With the goal of minimizing latency, the optimization variable is the offloading ratio of each device. In order to make the optimal offloading decision, the total latency function is first formulated. Considering that the local, MEC server and idle vehicle processing tasks are parallel, the total latency function is:

[0047]

[0048] Therefore, the optimization problem expression is:

[0049]

[0050] st C1:k0+k1+k2=1

[0051] C2:

[0052] C3:

[0053] C4:

[0054] Among them, v represents the vehicle's driving speed; constraint C1 means that the weighted sum of the allocation ratios is 1; constraint C2 means that the local computing amount cannot exceed the maximum computing amount; constraints C3 and C4 mean that the uploaded data amount must not exceed the maximum data amount.

[0055] Further, in step S6, solving the optimization model:

[0056] A latency-optimized task offloading and resource allocation parallel algorithm for in-vehicle edge computing environments is used to solve the optimization problem. The shortest latency load balancing algorithm aims to evenly distribute tasks to different execution units to achieve simultaneous completion of tasks and reduce the overall execution time. In this case, the task is divided into three parts, which are processed locally by the target vehicle, processed by the MEC server, and processed by the idle vehicle SV. At this time, when the latency of these three parts is equal, k0, k1, and k2 can be obtained respectively:

[0057]

[0058] in,

[0059] The present invention also discloses a parallel system for delay optimization task offloading and resource allocation for an on-board edge computing environment, which is used to execute the above method and includes the following modules:

[0060] Model building module: build a model for collaborative calculation between target vehicles and idle vehicles under the coverage of roadside unit RSU;

[0061] Initial system configuration module: configure the motion model of the target vehicle and idle vehicles, the calculation and unloading options of the target vehicle, the task allocation ratio and the communication range;

[0062] Communication model analysis module: provides the transmission rate formula and obtains the task transmission delay;

[0063] Computational model analysis module: calculation task delay;

[0064] Optimization model building module: with the goal of minimizing the delay of the parallel system, an optimization model for joint task allocation and delay minimization is established based on the task allocation ratio constraint and the delay constraint of vehicle task calculation offloading;

[0065] Optimization model solving module: The task allocation ratio is solved by making the delay of local processing of the target vehicle, the delay of processing by the mobile edge computing server equipped with the roadside unit, and the delay of processing by the idle vehicle equal.

[0066] The present invention has the following advantages:

[0067] (1) The present invention provides a parallel method and system for latency-optimized task offloading and resource allocation in a vehicle-mounted edge computing environment. By optimizing the task allocation ratio and making full use of a variety of computing resources, the total latency of the target vehicle to complete the task is significantly reduced.

[0068] (2) In intelligent transportation and IoT applications, real-time performance is crucial. The present invention uses mobile vehicles as edge computing nodes. By considering the dynamic mobility of vehicles, tasks can be better scheduled to ensure that they are unloaded and completed in time at the target vehicle, reducing latency. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 The present invention is a flowchart of a method for parallel delay-optimized task offloading and resource allocation for an on-vehicle edge computing environment according to a preferred embodiment of the present invention.

[0070] Figure 2 The relationship between task size and latency under different resource allocation schemes.

[0071] Figure 3 This is a graph showing the relationship between the local computing power and latency of the target vehicle under different resource allocation schemes.

[0072] Figure 4 This is a diagram showing the relationship between the computing power and latency of MEC servers under different resource allocation schemes.

[0073] Figure 5 This is a diagram showing the relationship between the computing power and latency of idle vehicle SVs under different resource allocation schemes.

[0074] Figure 6 This is a graph showing the relationship between the target vehicle’s transmission power and delay under different resource allocation schemes.

[0075] Figure 7 This is a block diagram of a parallel system for latency optimization task offloading and resource allocation for an on-board edge computing environment in a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0076] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0077] like Figure 1 As shown, this embodiment provides a parallel method for delay optimization task offloading and resource allocation for an in-vehicle edge computing environment, including the following steps:

[0078] S1. Establish a model: Establish a model for the target vehicle and idle vehicles to perform collaborative calculations under the coverage of a roadside unit RSU. The model is as follows: This embodiment considers a one-way straight road, along which a roadside unit is deployed. In addition to the target vehicle, there is also an idle friendly assisting vehicle in the system, and they are all traveling in the same direction under the RSU.

[0079] S2. Initial system configuration: The communication coverage of RSU is regarded as a cell, and RSU is equipped with a fixed mobile edge computing server, which can help the target vehicle perform its computing tasks. Assume that the vehicle travels at a constant speed. Consider that the target vehicle starts from the left end of the RSU coverage area and drives out of the rightmost end of the coverage area. The radius of RSU is R. At the junction of the road and the RSU coverage area, the angle between the road and the RSU radius is θ. Therefore, the distance traveled by the vehicle under RSU is expressed as:

[0080] K=2Rcosθ (1)

[0081] In order to complete the task of local calculation of the target vehicle before leaving the RSU, the constraint can be expressed as:

[0082]

[0083] In order to ensure that the transmission process of the target vehicle offloading to the MEC server of the RSU is completed before leaving the RSU, the constraint can be expressed as:

[0084]

[0085] In order to complete the transfer process from the target vehicle to the idle vehicle SV before leaving the RSU, the constraint can be expressed as:

[0086]

[0087] The task generated on the target vehicle is represented by <D, F>, where D represents the input data size and F represents the CPU cycles required to complete the task. The target vehicle can choose to perform its task locally or offload it to the MEC server of the RSU and execute it in parallel. The proportion of tasks assigned to each device is recorded as A = {k0, k1, k2}. k0, k1, k2 represent the proportion of tasks calculated locally, offloaded to the RSU (MEC server) and offloaded to idle vehicles, respectively. The proportion of these three parts is weighted to 1, which can be expressed as:

[0088] k0+k1+k2=1 (5)

[0089] According to the allocation ratio, the task size allocated to local computing and the number of CPU cycles required to complete this task can be expressed as:

[0090]

[0091] The task size assigned to the MEC server on the RSU and the number of CPU cycles required to complete this task are expressed as:

[0092]

[0093] The size of the task assigned to the idle vehicle and the number of CPU cycles required to complete this task are expressed as:

[0094]

[0095] S3. Analyze the communication model: At the physical layer, adaptive modulation and coding are assumed and adopted to support multiple transmission modes. In this system, the uplink and downlink channels between the target vehicle and the road facility unit and the idle vehicle are assumed to be frequency-flat block-fading Rayleigh channels. In order to transmit data, this embodiment adopts orthogonal frequency division multiplexing technology to ensure the orthogonality between different communication links and effectively reduce the impact of interference. Therefore, the data transmission rate from the target vehicle to the roadside unit RSU can be expressed as:

[0096]

[0097] The data transmission rate between the target vehicle and the idle vehicle SV can be expressed as:

[0098]

[0099] Among them, B M , B V2V They represent the channel bandwidth of the uplink between the target vehicle and the RSU (MEC server) and the idle vehicle SV, respectively, P represents the transmission power of the target vehicle device data, and d TM d V2V They represent the distances between the target vehicle and the RSU (MEC server) and between the target vehicle and the idle vehicle, ζ is the path loss index, and h M and h V2V Respectively represent the channel gain between the target vehicle and the RSU (MEC server) and the idle vehicle SV, and N0 is the additive white noise power. For the convenience of calculation, R is used to represent d TM , which can be expressed as:

[0100] d TM =R (11)

[0101] Therefore, the transmission delay of the task generated by the target vehicle to be uploaded to the RSU (MEC server) is:

[0102]

[0103] The transmission delay of uploading the task to the idle vehicle SV is:

[0104]

[0105] S4. Analytical calculation model, specifically:

[0106] When the target vehicle has a computing task, the computing task can be split to minimize the delay in completing the task. Therefore, the computing task <D, F> generated on the target vehicle can be simultaneously assigned to the local vehicle, the MEC server on the RSU, and the idle vehicle for parallel processing. The corresponding computing delay for local computing is:

[0107]

[0108] The task calculation delay transmitted to the RSU (MEC server) is:

[0109]

[0110] The task calculation delay transmitted to the idle vehicle part is:

[0111]

[0112] Among them, f L 、f M 、f S They represent the computing capabilities of the local MEC server, the RSU-equipped MEC server, and the idle vehicles, respectively.

[0113] S5. Establish an optimization model: Solve the task allocation ratio by making the local processing delay of the target vehicle, the processing delay of the mobile edge computing server equipped with the roadside unit, and the processing delay of the idle vehicle equal.

[0114] In order to make the best offloading decision with the best latency, we first formulate the total latency function. Considering that the local, MEC server and idle vehicle processing tasks are parallel, the total latency function is:

[0115]

[0116] Therefore, the optimization problem expression is:

[0117]

[0118] Among them, v represents the speed of the vehicle. Constraint C1 means that the weighted sum of the allocation ratio is 1; constraint C2 means that the local computing amount cannot exceed the maximum computing amount; constraint C3-4 means that the uploaded data amount cannot exceed the maximum data amount.

[0119] S6. Solve the optimization model, specifically:

[0120] A latency-optimized task offloading and resource allocation parallel algorithm for in-vehicle edge computing environments is used to solve the optimization problem. The shortest latency load balancing algorithm aims to evenly distribute tasks to different execution units to achieve simultaneous completion of tasks and reduce the overall execution time. In this case, the task is divided into three parts, which are processed locally by the target vehicle, processed by the MEC server, and processed by the idle vehicle SV. At this time, when the latency of these three parts is equal, k0, k1, and k2 can be obtained respectively:

[0121]

[0122] in,

[0123] Lemma: When When , the optimization target total delay T can be minimized.

[0124] Proof: Assume that at some point There are Consider the first-order derivative of T with respect to k0, k1, and k2 and

[0125] By definition, have:

[0126]

[0127] according to With the definition of , we can calculate the values ​​of these derivatives:

[0128]

[0129] Therefore, in At , these derivatives are all positive, that is, these three short times increase with the increase of k. However, k0+k1+k2=1, which means that at this time, as long as one of the three ratios decreases, the other will increase, and T will increase accordingly.

[0130] In summary, when And when k0+k1+k2=1, the total delay T is locally minimum.

[0131] The optimal local, RSU, V2V (LRV) collaboration strategy is obtained by the parallel method of latency-optimized task offloading and resource allocation for the vehicle-mounted edge computing environment proposed in the present invention. In the simulation analysis, in order to analyze the performance of the LRV allocation strategy, it is compared with four schemes. The four benchmark algorithms are: Scheme 1: strategy with only MEC server and idle vehicle SV auxiliary calculation; Scheme 2: strategy with only target vehicle local and MEC server calculation; Scheme 3: strategy with only target vehicle local and idle vehicle SV calculation; Scheme 4: strategy with only target vehicle local calculation.

[0132] Figure 2 The relationship between task size and latency under five schemes is plotted. It can be noted that as the task size increases, the latency of the V2V and RSU collaborative strategies gradually becomes greater than that of the local and RSU collaborative strategies. This is because when the task is small, the transmission latency and calculation latency offloaded to other devices will be greater than the latency directly calculated locally. However, as the task size increases, offloading to devices with stronger computing power such as MEC servers can make the total latency smaller. As can be seen from the figure, compared with the other four schemes, the method designed by the present invention has a lower task processing latency.

[0133] Figure 3 , 4 , 5 respectively plot the impact of different local computing power, different MEC computing power, different idle vehicle computing power, and different target vehicle transmission power on the total delay. Figure 3 , 4 , 5 It can be seen that when the local computing power, MEC computing power, and idle vehicle SV computing power are enhanced, the total delay of the collaborative strategy including the corresponding computing power will be correspondingly reduced. Compared with the other four schemes, the method designed by the present invention has a lower task processing delay in the same period.

[0134] Figure 6 The relationship between the target vehicle transmission power and latency under five schemes is shown. It can be seen that the increase in transmission power has almost no effect on the total latency, and the total latency is only slightly reduced. This is because the total latency is mainly determined by the processing time of the task, not just the transmission delay. Changing the transmission power does not directly affect the processing time of the task locally, on the idle vehicle SV or on the MEC server. Therefore, the impact of the change in transmission power on the total latency may be limited. Under limited influence, the method designed by the present invention still has the lowest latency.

[0135] like Figure 7 As shown, this embodiment discloses a parallel system for delay optimization task offloading and resource allocation for an in-vehicle edge computing environment, which is used to execute the above method embodiment and includes the following modules:

[0136] Model building module: build a model for collaborative calculation between target vehicles and idle vehicles under the coverage of roadside unit RSU;

[0137] Initial system configuration module: configure the motion model of the target vehicle and idle vehicles, the calculation and unloading options of the target vehicle, the task allocation ratio and the communication range;

[0138] Communication model analysis module: provides the transmission rate formula and obtains the task transmission delay;

[0139] Computational model analysis module: calculation task delay;

[0140] Optimization model building module: with the goal of minimizing the delay of the parallel system, an optimization model for joint task allocation and delay minimization is established based on the task allocation ratio constraint and the delay constraint of vehicle task calculation offloading;

[0141] Optimization model solving module: The task allocation ratio is solved by making the delay of local processing of the target vehicle, the delay of processing by the mobile edge computing server equipped with the roadside unit, and the delay of processing by the idle vehicle equal.

[0142] For other contents of this embodiment, please refer to the above method embodiment.

[0143] In summary, the resource allocation method designed in the present invention can significantly improve the utilization rate of device resources and make up for the deficiency of the existing edge computing resource allocation algorithm in offloading a single device.

[0144] The preferred embodiments and principles of the present invention are described in detail above. For those skilled in the art, according to the ideas provided by the present invention, there may be changes in the specific implementation methods, and these changes should also be regarded as the protection scope of the present invention.

Claims

1. A parallel method for latency optimization task offloading and resource allocation for an on-board edge computing environment, characterized by: The following steps are involved: S1. Establishing a model: Establishing a model for collaborative calculation between target vehicles and idle vehicles under the coverage of a roadside unit RSU; S2, initial system configuration: configure the motion model of the target vehicle and idle vehicles, the calculation and unloading options of the target vehicle, the task allocation ratio and the communication range; S3. Analyze the communication model: give the transmission rate formula and obtain the task transmission delay; S4, analytical computing model: task computing delay; S5. Establishing an optimization model: with the goal of minimizing the delay of the parallel system, an optimization model for joint task allocation and delay minimization is established according to the task allocation ratio constraint and the delay constraint of vehicle task calculation offloading; S6. Solve the optimization model: Solve the task allocation ratio by making the delay of local processing of the target vehicle, the delay of processing by the mobile edge computing server equipped with the roadside unit, and the delay of processing by the idle vehicle equal.

2. According to claim 1, a method for parallel delay optimization task offloading and resource allocation for an on-board edge computing environment is characterized in that: In step S1, the collaborative computing model is as follows: Suppose there is a one-way straight road with roadside units deployed along one side of the road. In addition to the target vehicle, there is also an assisting idle vehicle. They are all traveling in the same direction under the coverage of the RSU.

3. The method for parallel delay optimization task offloading and resource allocation for an on-board edge computing environment according to claim 1, characterized in that: In step S2, the initial system configuration is as follows: The vehicle's driving speed is constant, the communication coverage of the RSU is regarded as a cell, and the RSU is equipped with a fixed mobile edge computing server as an edge computing node; the target vehicle starts driving from the left end of the RSU coverage area until it drives out of the rightmost end of the coverage area; the radius of the RSU is R, and at the junction of the road and the RSU coverage area, the angle between the road and the RSU radius is θ, so the vehicle's driving distance under the RSU is expressed as: K=2Rcosθ The task generated on the target vehicle is represented by <D, F>, where D represents the input data size and F represents the CPU cycles required to complete the task. The target vehicle chooses to execute its task locally or offload it to the MEC server of the RSU for parallel execution. The proportion of tasks assigned to each device is recorded as A = {k0, k1, k2}, where k0, k1, and k2 represent the proportion of tasks calculated locally, offloaded to the RSU, and offloaded to idle vehicles, respectively. The proportions of these three parts are weighted as 1, expressed as: k0+k1+k2=1 According to the allocation ratio, the task size allocated to local computing and the number of CPU cycles required to complete this task are expressed as: The task size assigned to the MEC server on the RSU and the number of CPU cycles required to complete this task are expressed as: The size of the task assigned to the idle vehicle and the number of CPU cycles required to complete this task are expressed as:

4. According to claim 3, a method for parallel delay optimization task offloading and resource allocation for an on-board edge computing environment is characterized in that: In step S3, the communication model is as follows: Orthogonal frequency division multiplexing technology is used; the data transmission rate from the target vehicle to the roadside unit is expressed as: The data transmission rate between the target vehicle and the idle vehicle is expressed as: Among them, B M , B V2V represents the channel bandwidth of the uplink between the target vehicle, RSU and idle vehicle, P represents the transmission power of the target vehicle equipment data, d TM ,d V2V denote the distances between the target vehicle and the RSU and between the target vehicle and the idle vehicle, ζ is the path loss index, and h M 、h V2V They represent the channel gains of the target vehicle, RSU and idle vehicle respectively, N0 is the additive white noise power; R represents d TM , expressed as: d TM =R The transmission delay of the partial task generated by the target vehicle and uploaded to the RSU is expressed as: The transmission delay of uploading part of the tasks generated by the target vehicle to the idle vehicle is expressed as:

5. According to claim 4, a parallel method for delay optimization task offloading and resource allocation for an on-board edge computing environment is characterized in that: In step S4, the computing task 〈D, F〉 generated on the target vehicle is simultaneously assigned to the local vehicle, the MEC server on the RSU, and the idle vehicle for parallel processing; the corresponding computing delay of the local calculation is: The calculation delay of the task transmitted to RSU is: The task calculation delay transmitted to the idle vehicle is: Among them, f L 、f M 、f S They represent the computing capabilities of the target vehicle, the MEC server equipped by the RSU, and the idle vehicles respectively.

6. According to claim 5, a method for parallel delay optimization task offloading and resource allocation for an on-board edge computing environment is characterized in that: Step S5 is specifically as follows: With the goal of minimizing latency, the optimization variable is the offloading ratio of each device. The total latency function is formulated. Considering that the local, MEC server, and idle vehicle processing tasks are parallel, the total latency function is: Therefore, the optimization problem expression is: st C1:k0+k1+k2=1 Among them, v represents the vehicle's driving speed; constraint C1 means that the weighted sum of the allocation ratios is 1; constraint C2 means that the local computing amount cannot exceed the maximum computing amount; constraints C3 and C4 mean that the uploaded data amount must not exceed the maximum data amount.

7. The method of parallel delay optimization task offloading and resource allocation for vehicle-mounted edge computing environment according to claim 6, characterized in that: In step S6, the optimization model solving process is as follows: The task is divided into three parts, which are processed by the target vehicle locally, the MEC server, and the idle vehicle respectively. When the delays of these three parts are equal, k0, k1, and k2 are obtained as follows: in, 8. A latency-optimized task offloading and resource allocation parallel system for an in-vehicle edge computing environment, used to execute the method according to any one of claims 1 to 7, characterized in that: Includes the following modules: Model building module: build a model for collaborative calculation between target vehicles and idle vehicles under the coverage of roadside unit RSU; Initial system configuration module: configures the motion model of the target vehicle and idle vehicles, the calculation and unloading options of the target vehicle, the task allocation ratio and the communication range; Communication model analysis module: provides the transmission rate formula and obtains the task transmission delay; Computational model analysis module: calculation task delay; Optimization model building module: with the goal of minimizing the delay of the parallel system, an optimization model for joint task allocation and delay minimization is established based on the task allocation ratio constraint and the delay constraint of vehicle task calculation offloading; Optimization model solving module: The task allocation ratio is solved by making the delay of local processing of the target vehicle, the delay of processing by the mobile edge computing server equipped with the roadside unit, and the delay of processing by the idle vehicle equal.

Images