CN-116828536-B - Multi-target task unloading method and system for Internet of vehicles based on time delay perception energy conservation

CN116828536BCN 116828536 BCN116828536 BCN 116828536BCN-116828536-B

Abstract

The invention discloses a multi-target task unloading method and system of an Internet of vehicles based on time delay perception and energy conservation, wherein the method comprises the following steps of S1, obtaining information data, S2, calculating energy consumption and time delay, S3, determining a joint optimization problem of time delay and energy consumption, S4, determining an optimization objective function, S5, converging on a certain node of a solution space by utilizing an SA-PSO algorithm, bypassing a local optimal solution, so as to obtain a solution global optimal solution, S6, solving the joint optimization problem, dividing the solution into different types according to different task unloading positions, solving to obtain optimal unloading proportion and optimal transmitting power under different types, S7, selecting end equipment or corresponding node with the most residual energy, and selecting corresponding optimal unloading proportion and optimal transmitting power, and performing task unloading. The invention effectively shortens the task completion time and improves the task completion rate.

Inventors

LI PEI
YANG XU
LIANG XUESONG
WEI CHAO
YAO YINGBIAO

Assignees

杭州电子科技大学

Dates

Publication Date: 20260508
Application Date: 20230328

Claims (7)

1. A multi-target task unloading method of the internet of vehicles based on time delay perception energy conservation is characterized by comprising the following steps: S1, acquiring information data; S2, calculating energy consumption and time delay; S3, determining the joint optimization problem of time delay and energy consumption, specifically, in the step, introducing a time delay weight coefficient R T and an energy consumption weight coefficient R E by adopting a linear weighting sum method, setting the weight coefficients according to the state of the current vehicle-mounted equipment and task requirements and meeting R T +R E =1, defining an objective function: Wherein, the Local energy consumption for the end device; transmitting power for the end device; When the electric quantity of the automobile is sufficient, the time delay is a main optimization target, and the weight coefficient is set as When the electric quantity is insufficient, optimizing the energy consumption and setting ; The joint optimization problem of this step is described as follows: let the calculation rate of the fog node be u F , and the emission power of the fog node be The transmitting power of the cloud node is According to the definition of the energy consumption and the time delay function, the minimization problem of the time delay and the energy consumption function is expressed as: wherein, the unloading proportion of the four unloading modes of local processing, D2D, fog calculation and cloud calculation is respectively as follows Task average arrival rate of terminal equipment i The average computing power of the fog node i is The average task load rate of the end device i is The average transmission rate of the wireless port of the end device i is u D , and the sum of service request rates of the fog nodes is The method comprises the steps of (1) carrying out local calculation to ensure that the load capacity does not exceed the local calculation capacity, wherein p i is the user transmission power, (12-2) indicating that the task amount offloaded through D2D does not exceed the task amount born by a help node, (12-3) indicating that the sum of service request rates of the mist nodes is smaller than the calculation rate of the mist nodes, (12-4) indicating that the sum of service request rates of the mist nodes is smaller than the transmission power of the mist nodes, (12-5) indicating that the task amount offloaded to the cloud nodes is smaller than the calculation capacity of the cloud nodes, and (12-6) indicating that the transmission power of end equipment is not greater than the maximum transmission power, wherein the sum of the task offloaded proportion is between 0 and 1, and the sum of the service request rates offloaded to the help nodes and the local offloaded proportion is 1; S4, determining an optimization objective function; S5, utilizing an SA-PSO algorithm to converge on a certain node of a solution space and simultaneously bypassing a local optimal solution, so as to obtain a solution global optimal solution; s6, solving the joint optimization problem, dividing the joint optimization problem into different types according to different task unloading positions, and solving to obtain the optimal unloading proportion and the optimal transmitting power under different types; and S7, selecting the terminal equipment or the corresponding node with the largest residual energy, and selecting the corresponding optimal unloading proportion and the optimal transmitting power to carry out task unloading.
2. The method for unloading the internet of vehicles multi-target tasks based on time delay perception energy conservation as claimed in claim 1, wherein in step S1, the information data comprises the number N of the terminal devices, the number m of the fog node terminals and the average task data size of the terminal devices i End device i selects to offload the ratio The arrival rate of the terminal task is as follows Channel bandwidth B, signal to noise ratio Channel rate R i .
3. The method for unloading the internet of vehicles multi-target tasks based on time delay perception energy conservation as claimed in claim 2, wherein in step S2, the average time delay is calculated as follows: a) The transmission delay is that the user unloads the task to the cloud node, the fog node or transmits the task to other equipment nodes through D2D and is called as a help node, when the task is unloaded to the help node, only the uplink transmission delay of the user is considered : Wherein B represents the bandwidth of the channel, For signal-to-noise ratio, the signal-to-noise ratio is positively correlated with the transmission power, and the transmission delay is Inversely related to the channel rate R i , so the higher the transmit power, the more delay is Smaller, simultaneous transmission delay Ratio to unloading Is positively correlated, thus unloading the proportion The larger the transmission delay The larger; b) Local processing time delay, namely setting the average computing power of the terminal equipment i as And obtaining local calculation time delay T j1 of the end equipment: c) D2D processing time delay, namely combining an M/M/1 queuing theory in a queuing theory, wherein the average processing time of the D2D unloading task is as follows: d) Mist computation processing latency-allowing multiple end devices to offload tasks to a mist node layer, at which time the maximum request rate of the mist node Let the average computing power of the fog node i be The sum of service request rates of the fog nodes is: the sum of the time delays of the foggy nodes is: e) Cloud computing processing time delay, namely when the computing capacity of the fog node layer is insufficient to bear the current task unloading, the unloading proportion is as follows The task unloading to the cloud node with stronger computing capability is performed immediately because the cloud node and the cloud node are connected through the wired optical fiber to generate fixed time delay T 0 , and the task unloading of the task amount can be borne by the performance of a server in the cloud node, so the task unloading to the cloud node is performed immediately, the waiting time is ignored, and the cloud node execution process is regarded as the process in the queuing theory according to the analysis of the queuing theory The queue is provided with the calculation power u c of the cloud node, and the time delay of the cloud node layer is as follows: 。
4. the method for unloading the internet of vehicles multi-target tasks based on time delay perception energy conservation as claimed in claim 3, wherein in step S2, the calculated energy consumption is as follows: The energy consumption generated in the transmission process is related to the transmission power, load and transmission time, and the transmission power of the end equipment The method comprises the following steps: Local energy consumption of end devices The method comprises the following steps: Wherein q i is the unit operation power coefficient of the ith device and is a fixed constant; m jk is specifically as follows: the limitation is that for any row k of the matrix M, there is I.e. each task can only select one task offloading mode.
5. The method for unloading the internet of vehicles multi-target tasks based on time delay perception energy conservation as claimed in claim 4, wherein in step S5, a simulated annealing particle swarm algorithm, namely an SA-PSO algorithm, is described as follows: (1) Particle swarm algorithm Is arranged in the D-dimensional search space, a population consisting of M particles is subjected to 'flying', i.e. searching, and the position of the ith particle is expressed as The speed of the ith particle is expressed as The best position searched by each particle in the flying process is compared with the historical best fitness value of the particle, which is called an individual extremum The best place searched for by the whole population is called global extremum The speed and location update formula in the particle search process is as follows: Wherein i=1, 2, 3..n, d=1, 2, 3..d, w is an inertial weight factor, c 1 and c 2 are learning factors, also called acceleration factors, taking negative numbers, r 1 and r 2 are two random numbers between [0,1 ]; (2) Simulated annealing algorithm The simulated annealing algorithm needs to set an initial temperature according to the initial state of the population in the initial stage of iteration, judges whether a new solution generated by interference replaces a global optimal solution according to a Mitreboles criterion when the temperature of the internal particles of the simulated solid is reduced in each iteration, and has the following expression: Wherein E i (k) represents the internal energy of the ith particle in the kth iteration, namely the fitness value of the current particle, E g represents the internal energy of the optimal point of the current population, T i represents the current temperature, each iteration of the temperature is linearly attenuated to a certain degree, the optimizing process is an alternating process of continuously searching for a new solution and slowly reducing the temperature, and E i (k) determines the new state E i (k+1) generated next time and is irrelevant to the previous E i (0) to E i (k-1).
6. The method for multi-objective task offloading of the internet of vehicles based on time delay perception energy conservation as claimed in any one of claims 1 to 5, wherein in step S6, according to different task offloading positions, the method is divided into three types, namely task offloading of end equipment-end equipment, task offloading of end equipment-fog node, and task offloading of end equipment-fog node-cloud node.
7. A time delay aware energy efficient internet of vehicles multi-objective task offloading system based on the method of any one of claims 1-6, comprising the following modules: the information data acquisition module is used for acquiring information data; The calculation module is used for calculating energy consumption and time delay; the joint optimization problem module is used for determining a joint optimization problem of time delay and energy consumption; the optimization objective function module is used for determining an optimization objective function; the global optimal solution solving module is used for converging the SA-PSO algorithm to one node of the solution space and bypassing the local optimal solution at the same time so as to obtain a solution for solving the global optimal solution; The optimal unloading proportion and optimal transmitting power solving module is used for solving the joint optimization problem, dividing the joint optimization problem into different types according to different task unloading positions, and solving the joint optimization problem to obtain the optimal unloading proportion and the optimal transmitting power under different types; and the selecting and unloading module is used for selecting the end equipment or the corresponding node with the largest residual energy, selecting the corresponding optimal unloading proportion and the optimal transmitting power and unloading the task.

Description

Multi-target task unloading method and system for Internet of vehicles based on time delay perception energy conservation Technical Field The invention belongs to the technical field of task unloading of the Internet of vehicles. According to the method and the system for unloading the internet of vehicles multi-target tasks, different processing modes are selected according to different service types, a multi-target optimization problem related to time delay and energy consumption is modeled, equipment time delay and energy consumption are minimized by adjusting task unloading proportion and uplink terminal transmitting power, and particularly the method and the system for unloading the internet of vehicles multi-target tasks based on time delay perception and energy saving are related. Background Along with the exponential increase of the data volume required to be transmitted and processed in the internet of vehicles, the data collected by the vehicles is transmitted to a cloud computing platform to be processed, so that the load of the network transmission bandwidth is increased sharply, the network time delay is greatly increased, and the processing requirement of time delay sensitive tasks in the internet of vehicles cannot be met. Cloud convergence networks are considered to be a very potential network architecture and have gained wide acceptance in the industry, demonstrating further improvements in computing power, expansion of communication coverage and reduction of transmission delays. However, while the time delay performance of the user is improved, more energy consumption is likely to be brought about, and the shortage of energy is becoming a key obstacle for limiting the development of the internet of vehicles. In particular, electric vehicles or hybrid electric vehicles will dominate the market in the near future, which has prompted more attention to energy consumption problems in the internet of vehicles. In real life, each vehicle typically requires different computing resources due to the different size of the computing tasks. During the unloading of computing tasks, limited computing resources are shared among vehicles, each of which needs to determine whether to unload. Thus, minimizing average costs within the system by jointly optimizing task offloading decisions is intuitively important to improve overall system performance. The current increase in the number of connected devices is accompanied by a rapid increase in network traffic and data center load, which results in a significant increase in power consumption of the communication network and data center, while at the same time creating a significant challenge for low latency and highly reliable data forwarding requirements in vehicle communications. With the progress of information technology, delay-sensitive application layers such as road environment augmented reality, man-vehicle dynamic interaction, driving safety early warning and the like are endless, wireless network access is frequently changed, system resources are dynamically allocated, nodes are rapidly changed, and the requirements on low-delay high-reliability data forwarding in vehicle-mounted communication are more strict. Disclosure of Invention Aiming at the problem of low power consumption of delay sensitive service in the prior art, the invention provides a multi-target task unloading method and system for the Internet of vehicles based on delay sensing and energy saving. The invention designs a network architecture and a task unloading flow based on vehicle collaborative edge calculation, analyzes the influence of communication resource allocation and calculation resource allocation on task completion time, builds a task completion time model, builds an optimization model with the aim of minimizing average completion time and maximizing successful completion rate, and designs a particle swarm algorithm and an annealing algorithm based on the optimization model for solving. Simulation results show that compared with strategies adopted in the prior art, the method effectively shortens the task completion time and improves the task completion rate by using a simulated annealing particle swarm algorithm (Simulated Annealing-PARTICLE SWARM Optimization, SA-PSO). In order to achieve the above object, the present invention adopts the following scheme: A multi-target task unloading method of the Internet of vehicles based on time delay perception energy conservation comprises the following steps: S1, acquiring information data; Preferably, the information data comprises the number N of end devices, the number m of fog node ends, the average task data size theta i of the equipment i, the average task arrival rate lambda i of the end devices i, the terminal i selects the unloading proportion beta i(0≤βi less than or equal to 1), and the terminal task arrival rate is Local processing, D2D, fog calculation and cloud calculation, wherein the unloading proportion of the fo