CN113590211A - Calculation unloading method based on PSO-DE algorithm - Google Patents
Calculation unloading method based on PSO-DE algorithm Download PDFInfo
- Publication number
- CN113590211A CN113590211A CN202110532059.5A CN202110532059A CN113590211A CN 113590211 A CN113590211 A CN 113590211A CN 202110532059 A CN202110532059 A CN 202110532059A CN 113590211 A CN113590211 A CN 113590211A
- Authority
- CN
- China
- Prior art keywords
- task
- particle
- population
- unloading
- algorithm
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000004364 calculation method Methods 0.000 title claims abstract description 18
- 239000002245 particle Substances 0.000 claims abstract description 88
- 238000005457 optimization Methods 0.000 claims abstract description 31
- 238000005265 energy consumption Methods 0.000 claims abstract description 28
- 230000035772 mutation Effects 0.000 claims abstract description 16
- 230000006870 function Effects 0.000 claims description 14
- 101150000810 BVES gene Proteins 0.000 claims description 13
- 101150025129 POP1 gene Proteins 0.000 claims description 13
- 230000003044 adaptive effect Effects 0.000 claims description 13
- 238000012545 processing Methods 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 6
- 101150030531 POP3 gene Proteins 0.000 claims description 5
- 230000000694 effects Effects 0.000 claims description 4
- 101100273030 Schizosaccharomyces pombe (strain 972 / ATCC 24843) caf1 gene Proteins 0.000 claims description 3
- 230000009286 beneficial effect Effects 0.000 claims description 3
- 230000005540 biological transmission Effects 0.000 claims description 3
- 230000009191 jumping Effects 0.000 claims description 3
- 238000013459 approach Methods 0.000 claims description 2
- 238000002474 experimental method Methods 0.000 claims description 2
- 238000007689 inspection Methods 0.000 claims description 2
- 230000002028 premature Effects 0.000 claims description 2
- 230000002708 enhancing effect Effects 0.000 claims 1
- 238000007670 refining Methods 0.000 claims 1
- 230000003595 spectral effect Effects 0.000 claims 1
- 238000011160 research Methods 0.000 description 4
- 230000001174 ascending effect Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000002431 foraging effect Effects 0.000 description 2
- 238000013468 resource allocation Methods 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/445—Program loading or initiating
- G06F9/44594—Unloading
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a calculation unloading method based on a PSO-DE algorithm, which mixes the PSO algorithm and the DE algorithm. The DE algorithm is introduced, diversity of unloading strategy combination is increased and accuracy of finding the optimal unloading strategy is improved by improving operations such as mutation, intersection and the like, and the self-adaptive mutation operator is combined, so that the global and local optimization capability of the particle swarm algorithm is favorably balanced, and meanwhile, the particle swarm algorithm searching efficiency and the optimization accuracy are improved. The invention converts the problem of finding the optimal unloading strategy into the problem of finding the optimal position of the particles, and each particle represents a task unloading scheme. The method has stronger searching capability and optimizing capability on the aspect of searching the optimal task unloading strategy, can find the unloading strategy with the lowest energy consumption from a plurality of task unloading schemes, ensures the service quality of a user to a certain extent, improves the endurance of equipment and improves the performance of the mobile terminal.
Description
Technical Field
The invention relates to the field of mobile edge calculation, in particular to a calculation unloading method based on a PSO-DE algorithm
Background
With the rise of 5G networks and the continuous development of wireless networks, Mobile Devices (MDs) are becoming an irreplaceable part of people's lives. Meanwhile, the living demands of people are continuously improved, and a large number of high-quality services and applications, such as smart home, VR, human face and fingerprint identification, are promoted. These applications typically result in high energy consumption and require powerful computing power, but such tasks may not be efficiently and timely handled due to insufficient computing power of the mobile terminal device's own memory space. Although the advent of Mobile Cloud Computing (MCC) can provide the ability to handle complex Computing and high storage for devices with insufficient storage space and Computing power, it still has some disadvantages: cloud computing is centralized data processing, that is, each task needs to be transmitted from a terminal to a remote cloud terminal through a remote network. When the task amount is continuously increased, the cloud server is far away from the terminal equipment, so a series of problems such as data loss, high energy consumption, network delay and the like are easily caused in the data transmission process, and the user experience is seriously influenced by the problems.
To further reduce system energy consumption and time delay, improve user experience quality, etc., Mobile Edge Computing (MEC) is introduced. The moving edge calculation is a further optimization of cloud calculation: the reasonable computation and unloading strategy is adopted to download the computation storage and the computation resources to the edge of the network close to the mobile equipment, so that the problems of high delay, high energy consumption, network delay, data leakage and the like caused by centralized cloud computing are solved, the computation and storage pressure of the terminal equipment is relieved, and the service life of the equipment is prolonged. In recent years, computational offloading has become a research hotspot in the field of mobile edge computing, and since 2014, many researchers have begun to research and obtain many research results. The computational offloading mainly includes two aspects: offloading decisions and resource allocation. The unloading decision is mainly researched on how to divide tasks generated by terminal equipment and unload part of tasks to the MEC server for execution by measuring benefits generated by strategies, namely, how to reasonably allocate limited MEC server resources to each task is a very critical problem. Therefore, it is necessary to provide an efficient and reasonable computation offload strategy.
At present, the most widely studied is to take the time delay as a main optimization target, which often neglects that too much energy consumption also affects the performance and service life of the terminal equipment, and also affects the user experience. Aiming at the defects of the energy optimization, a scene of a multi-user single MEC server is constructed (see figure 1), and considering that the resource of the single MEC server is limited, how to reduce the total energy consumption of task execution through a reasonable task unloading strategy and a resource allocation technology is a key of research, and the improvement of the terminal endurance is realized. A calculation unloading model of a multi-user single MEC server comprehensively considering time delay and energy consumption is provided, namely an unloading strategy problem of a plurality of tasks is converted into a problem of solving the minimum energy consumption of all tasks, and a penalty function is used for balancing the problem between the time delay and the energy consumption in consideration of the fact that the time delay is also an important factor influencing the user experience quality. The task is determined to be executed locally or on an MEC server by comparing the consumption conditions of the task at different positions, and finally the optimal unloading decision is found. However, as the amount of tasks increases, the calculation process is too complex and the number of combinations of the unloading strategies increases exponentially, and it is almost impossible to achieve the optimal unloading strategies by directly using an enumeration method, but the heuristic algorithm has the advantages that it is more efficient than blind search, and if a heuristic algorithm is carefully improved, the optimal solution of the optimization problem can be obtained in a short time.
The Particle Swarm Optimization (PSO) core idea is to simulate the behavior of a bird Swarm searching food in an unknown space, and to enable the movement of the whole Swarm to generate an evolution process from disorder to order in a problem solving space through mutual cooperation and information exchange among individuals in the Swarm, so as to obtain the optimal solution of the problem. The Differential Evolution (DE) algorithm was first proposed by Storm et al in 1995 and is an efficient global optimization algorithm. The population is updated by carrying out operations such as variation, crossing, selection and the like on individuals according to difference information among the individuals in the initialized population, and in the iterative process, an elite population is reserved according to a rule of 'winning or losing priority', and the whole optimization process is guided to gradually approach to the optimal solution.
The two intelligent algorithms have good global optimization capability, high search speed and easy realization. However, as the number of tasks increases, the later convergence speed of the algorithms is slow, and the algorithms are easy to fall into local optimization, so that the optimization precision of the algorithms is reduced, and large-scale task unloading can not be basically met.
Most of the prior heuristic algorithms still have some problems in solving the problem of finding the optimal task unloading strategy in the mobile edge calculation, cannot ensure the optimization precision, is easy to fall into the local optimum and the like. In the invention, a difference DE algorithm is introduced into a PSO algorithm, the diversity of unloading strategy combination is increased and the accuracy of searching for an optimal unloading strategy is improved by improving operations such as variation, intersection and the like, and the self-adaptive mutation operator is combined, so that the global and local optimization searching capability of the particle swarm algorithm is favorably balanced, and meanwhile, the searching efficiency and the optimization searching accuracy of the particle swarm algorithm are also improved.
Disclosure of Invention
The invention provides a calculation unloading method of a PSO-DE algorithm based on an MEC system (as shown in figure 1) of a multi-user single MEC server, and the method introduces the DE algorithm on the basis of the PSO algorithm, increases the diversity of unloading strategy combinations and improves the accuracy of searching for an optimal unloading strategy by improving the operations of variation, intersection and the like; and the self-adaptive mutation operator is combined, so that the global and local optimization capability of the particle swarm optimization is balanced, and the searching efficiency and the optimization accuracy of the particle swarm optimization are improved. As shown in FIG. 2, the PSO-DE algorithm-based computational offload strategy disclosed by the invention comprises the following steps:
step 1: initialization: randomly generating a population pop0 according to the dimension (user task number) of the particle individual, the particle code, the population scale and the like; initializing a learning factor, a maximum iteration number, a variation factor, a cross probability and the like;
step 2: according to the initialized population pop0, calculating the fitness value of each particle according to the fitness function and recording the global optimal position and the corresponding fitness value; note that each individual particle is now best initialized for the position of the particle.
And step 3: updating the speed and the position of each dimension of the particle according to a formula of the speed and the position of the particle, and performing border crossing processing on the speed and the position of each dimension of the particle;
and 4, step 4: and (4) selecting. And calculating the adaptive value of the updated particle according to the fitness function, and comparing the adaptive value with the adaptive value of the original particle individual. If the fitness value of the particle is larger than that of the original particle individual, replacing the original particle individual with a new particle, finally forming a new population pop1 and calculating the fitness of each particle individual of the population;
and 5: and (5) carrying out mutation. Carrying out mutation operation on the particle individuals in the population pop1 according to a mutation operator, and carrying out boundary inspection on each dimension of a new particle individual to finally form a population pop 2;
step 6: and (4) crossing. Performing cross operation on the individuals of the current population pop1 and the individuals of the varied population pop2, performing boundary check on each dimension of the new particle individuals, and finally forming a population pop3 and calculating the fitness of each particle individual of the population;
and 7: and (4) selecting. And sorting the fitness values of the particle individuals in the population pops 1 and 3 in ascending order, selecting the individuals with lower fitness values as the initial population pops 1 of the next iteration, and updating the optimal individuals, the global optimal values and the fitness values of the optimal individuals and the global optimal values respectively.
And 8: and judging whether the ending condition is met. If not, directly jumping to the step 3, and repeatedly executing the steps 3-7; otherwise, the loop is exited, and the optimal task unloading strategy and the corresponding adaptive value are directly output.
The invention has the beneficial effects that: the invention fuses the PSO algorithm and the DE algorithm and improves the inertia weight and the variation operation of the PSO algorithm and the DE algorithm respectively, thereby not only balancing the global and local optimizing capability of the population, but also avoiding premature convergence in the early stage and falling into the local optimal solution. The experimental results prove that: the PSO-DE calculation unloading strategy optimizing capability provided by the invention is superior to that of the traditional heuristic algorithm.
Drawings
FIG. 1 is an MEC system model
FIG. 2 is a flow chart of a PSO-DE algorithm-based calculation unloading method of the present invention
Detailed Description
The invention provides a calculation unloading method based on a PSO-DE algorithm, which is explained in detail below with reference to the accompanying drawings.
As shown in FIG. 2, the invention provides a calculation unloading method based on a PSO-DE algorithm, which comprises the following steps:
step 1: initialization: randomly generating pop0 according to the dimension N (user task number) of the particle individual, the particle code and the population model P; initializing learning factors c1, c2, maximum iteration number tmaxMutation factor F, crossover probability CR, etc.;
step 2: according to the initialized population pop0, calculating the fitness value of each particle according to the fitness function and recording the global optimal position gbestAnd its corresponding fitness value; note that at this point, each particle individually optimizes the position p initialized for that particlebesti。
In the invention, binary coding is adopted to describe the decision of task unloading, if s isiIf 0, the task needs to be executed locally; otherwise, the task needs to be offloaded to the MEC server for execution. Assume that there are currently k tasks, i.e. there are 2 in totalkA task offloading policy, using S ═ S1，...，si，...，skRepresents one of the task unloading schemes (i.e. a particle), and converts the problem of finding the best task unloading strategy into the problem of solving the optimal solution of the vector S, SiRepresents the location where task i was performed and siE {0, 1 }. Table 1 is the relationship between PSO-DE particle foraging and task unloading strategy:
TABLE 1 relationship of particle foraging Process to task offloading strategies
Each particle represents a feasible scheme of the current task unloading, and the degree of goodness of the task unloading strategy corresponding to the particle is measured by calculating the fitness value of each particle. The optimization problem of the invention is to minimize the total energy consumed by the task under the constraint of the maximum tolerance time of the user.
When user task i is executed on the own device, si0 is the time delay required to perform the task
And energy consumption
Are respectively represented as
When task i is offloaded to the MEC server for execution, si1 is the time delay required to perform the task
And energy consumption
Are respectively represented as
The optimization objectives of the present invention are: how to provide a user with a good service and simultaneously minimize the energy consumption for completing the task execution, thereby obtaining an optimal unloading strategy. Considering the problem of high delay, in order to avoid the phenomenon of low energy consumption replacing high delay, a penalty function is introduced: if the execution time of the task exceeds the maximum time delay which can be tolerated by the user, increasing the total energy consumption of the unloading strategy through a penalty function serves as a means for penalizing the unloading centralization of the task. Based on the above formula and considering the multi-objective constraint problem, the optimization objective can be defined as:
each particle represents a feasible scheme of the current task unloading, and the degree of goodness of the task unloading strategy corresponding to the particle is measured by calculating the fitness value of each particle. The optimization problem of the invention is to minimize the total energy consumed by the task under the constraint of the maximum tolerance time of the user, so the fitness value of the particle is mainly influenced by two factors of the execution delay and the energy consumption of the task. The greater the energy consumption for completing the task, the poorer the optimization effect; the smaller the energy consumption is, the better the optimization effect is; based on the above formula, the fitness function may be constructed as
And step 3: respectively updating the speed v of each dimension of the particles according to the following formulaijAnd position xijAnd performing border crossing processing on the speed and the position of each dimension of the particle;
and 4, step 4: and (4) selecting. And calculating the adaptive value of the updated particle according to the fitness function, and comparing the adaptive value with the adaptive value of the original particle individual. If the fitness value of the particle is larger than that of the original particle individual, replacing the original particle individual with a new particle, finally forming a new population pop1 and calculating the fitness of each particle individual of the population;
and 5: and (5) carrying out mutation. Pairing particles in the population pop1 according to mutation operators
The individual is subjected to mutation operation, and new particle individuals are required
Performing boundary check on each dimension of the group to finally form a population pop 2;
F＝F*2δ
wherein, alpha and beta are respectively between [0, 1 ]]And α + β ═ 1; i ≠ r1 ≠ r2 ≠ r3, and i, r1, r2, r3 ∈ [1, P]∩Z；F0For mutation operators, F ∈ [ F ∈ ]0，2F0]。
Step 6: and (4) crossing. For individuals of the current population pop1
And the individuals of the varied population pop2
Performing crossover operation and performing new particle individuals
Performing boundary check on each dimension of the population, finally forming a population pop3 and calculating the fitness of each particle individual of the population;
wherein: j is an element of {1, 2.,. n }, and rand is an element ofIn [0, 1 ]]Is a cross probability and CR ∈ (0, 1), jrandE {1, 2.., n }, which is a random integer.
And 7: and (4) selecting. And sorting the fitness values of the particle individuals in the population pops 1 and 3 in ascending order, selecting P individuals with lower fitness values as initial population pops 1 of the next iteration, and updating the optimal individuals, the global optimal values and the fitness values of the optimal individuals and the global optimal values respectively.
And 8: and judging whether the ending condition is met. If not, directly jumping to the step 3, and repeatedly executing the steps 3-7; otherwise, the loop is exited, and the optimal task unloading strategy and the corresponding adaptive value are directly output. The following describes the experimental procedures and results of the present invention.
To verify the performance of the algorithm, first we compare the performance of the algorithm proposed by the present invention with the following 5 task offloading schemes: a local offload policy (all-local) is to process all tasks on the device itself; an MEC unloading strategy (all-EMC) is used for unloading all tasks to an MEC server for execution; a particle swarm algorithm-based offload strategy (PSOA), i.e., a traditional particle swarm algorithm; an unloading strategy (DE) based on a differential evolution algorithm, namely a traditional differential evolution algorithm; the offloading strategy (LAWPSO) of the PSO algorithm based on the linear self-decreasing weight is a particle swarm algorithm introducing the adaptive inertial weight. In contrast, the PSO-DE algorithm provided by the invention has stronger searching capability and optimizing capability.
The above description is only of the preferred embodiments of the present invention, and it should be noted that: it will be apparent to those skilled in the art that various modifications and adaptations can be made without departing from the principles of the invention and these are intended to be within the scope of the invention.
Claims (8)
1. The calculation unloading method based on the PSO-DE algorithm is characterized by comprising the following steps of:
step 1: initialization: randomly generating a population pop1 according to the dimension (user task number) of the particle individual, the particle code, the population scale and the like; initializing a learning factor, a maximum iteration number, a variation factor, a cross probability and the like;
step 2: according to the initialized population pop0, calculating the fitness value of each particle according to the fitness function and recording the global optimal position and the corresponding fitness value; note that each particle individual is now the best position to initialize for that particle;
and step 3: updating the speed and position of each dimension of the particle according to a formula of the speed and position of the particle, and performing border crossing processing on the speed and position of each dimension of the particle;
and 4, step 4: and (4) selecting. And calculating the adaptive value of the updated particle according to the fitness function, and comparing the adaptive value with the adaptive value of the original particle individual. If the fitness value of the particle is larger than that of the original particle individual, replacing the original particle individual with a new particle, finally forming a new population pop1 and calculating the fitness of each particle individual of the population;
and 5: and (5) carrying out mutation. Carrying out mutation operation on the particle individuals in the population pop1 according to a mutation operator, and carrying out boundary inspection on each dimension of a new particle individual to finally form a population pop 2;
step 6: and (4) crossing. Performing cross operation on the individuals of the current population pop1 and the individuals of the varied population pop2, performing boundary check on each dimension of the new particle individuals, and finally forming a population pop3 and calculating the fitness of each particle individual of the population;
and 7: and (4) selecting. And sorting the fitness values of the particle individuals in the populations pop1 and pop3 in a descending order, selecting the individuals with higher fitness values as the initial population pop1 of the next iteration, and updating the optimal individuals, the global optimal values and the fitness values of the optimal individuals and the global optimal values respectively.
And 8: and judging whether the ending condition is met. If not, directly jumping to the step 3, and repeatedly executing the steps 3-7; otherwise, the loop is exited, and the optimal task unloading strategy and the corresponding adaptive value are directly output.
2. The PSO-DE algorithm-based computational offload combination method of claim 1, wherein: in step 1, the present invention is considered in an MEC system consisting of U mobile devices and a base station. The system is mainly divided into two parts: (1) the mobile equipment is mainly responsible for processing tasks of the mobile equipment and transmitting unloading task data; (2) the MEC server, consisting of a base station and a server, is responsible for receiving and processing the tasks transmitted from the mobile devices. Wherein, each edge server can serve a plurality of mobile devices, and the task generated by each mobile device can be executed on the own device or the wireless network is connected to unload the task to the edge server. Note that: when the task unloading occurs, the invention only considers the processing by one edge server, and does not consider the partial unloading problem, such as the task executed on the MEC server is unloaded to the terminal equipment for processing in the middle.
3. The PSO-DE algorithm-based computational offload combination method of claim 1, wherein: in the step 1, as with the conventional group intelligent optimization algorithm, the algorithm provided by the invention still needs to determine the learning factor, the inertia weight, the mutation operator and the crossover probability according to experiments.
4. The PSO-DE algorithm-based computational offload method according to claim 1, characterized in that: in the step 2, the fitness of each particle individual is calculated in the following manner: the invention provides a calculation unloading model of a multi-user single MEC server, which comprehensively considers time delay and energy consumption, namely, an unloading strategy problem of a plurality of tasks is converted into a problem of solving the minimum energy consumption of all tasks, and a penalty function is used for balancing the problem between the time delay and the energy consumption in consideration of the fact that the time delay is also an important factor influencing the user experience quality. The invention adopts binary coding to describe the decision of task unloading, if siIf 0, the task needs to be executed locally; otherwise, the task needs to be offloaded to the MEC server for execution. Assume that there are currently k tasks, i.e. there are 2 in totalkA task offloading policy, using S ═ S1，...，si，...，skRepresents one of the task unloading schemes (i.e. a particle), and the problem of finding the best task unloading strategy is transferred toTo solve the problem of the optimal solution of the vector S, SiRepresents the location where task i was performed and si∈{0，1}；
When s isiWhen the local device is not executing the task i, the local device executes the task i; the time delay required to perform the task
And energy consumption
Are respectively represented as
Wherein D isiSize of input data amount for task i, ciIs the CPU cycle required to process one bit of data,
is the computing power of the device, k is a constant related to the chip structure of the terminal device;
when s isiWhen the task i is equal to 1, unloading the task i to an MEC server for execution; the time delay required to perform the task
And energy consumption
Are respectively represented as
Transmission rate for task i to transmit task data to MEC server
Wherein, W, (l)0)-σ、N0Respectively, the channel bandwidth, the channel gain and the noise power spectral density of the base station for data transmission between the terminal equipment and the base station; liAnd sigma are the distance between the user and the base station and the path loss factor respectively.
The optimization objectives of the present invention are: the optimal unloading strategy is obtained by minimizing the energy consumption of task execution completion while providing the user with high-quality service. Considering the problem of high delay, in order to avoid the phenomenon of low energy consumption replacing high delay, a penalty function is introduced: when the time for executing the task exceeds the maximum time delay which can be tolerated by the user, the total energy consumption of the unloading strategy is increased through a penalty function as a means for punishing the unloading centralization of the task. Based on the above formula and considering the multi-objective constraint problem, the optimization objective can be defined as:
wherein
。
5. The fitness calculation method according to claim 4, characterized in that: each particle represents a feasible scheme of the current task unloading, and the degree of the task unloading strategy corresponding to each particle is measured by calculating the fitness value of each particle. The greater the energy consumption for completing the task, the poorer the optimization effect; the smaller the energy consumption, the better the optimization effect. The fitness function of the particle is
6. The PSO-DE algorithm-based computational offload method according to claim 1, characterized in that: in the step 5, the population variation adopts a differential variation mode to obtain a new population, and the global and local optimizing capability of the population is balanced. In order to prevent phenomena such as premature convergence, the variation factor F is changed into a self-adaptive variation operator, the F value is very large at the initial stage of population evolution, and at the moment, the population needs to expand the range to search for a global optimal solution, so that the global search capability is improved; as the number of iterations increases, F gradually approaches F0The method is beneficial to refining the search range and enhancing the local optimization capability.
7. The PSO-DE algorithm-based computational offload method according to claim 1, characterized in that: in the step 6, the newly updated population pop1 and the variant population pop2 are subjected to cross operation according to the cross probability, so that the global optimizing capability is effectively enhanced, the population diversity is improved, and the searching capability of the algorithm is improved.
8. The PSO-DE algorithm-based computational offload method according to claim 1, characterized in that: in the step 7, a greedy strategy is adopted to select the particle individuals with better fitness from the pop1 and the pop3 in the population after crossing and mutation as the initial population of the next iteration, which is beneficial to accelerating the convergence speed of the population.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110532059.5A CN113590211A (en) | 2021-05-14 | 2021-05-14 | Calculation unloading method based on PSO-DE algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110532059.5A CN113590211A (en) | 2021-05-14 | 2021-05-14 | Calculation unloading method based on PSO-DE algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113590211A true CN113590211A (en) | 2021-11-02 |
Family
ID=78243169
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110532059.5A Pending CN113590211A (en) | 2021-05-14 | 2021-05-14 | Calculation unloading method based on PSO-DE algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113590211A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114143814A (en) * | 2021-12-13 | 2022-03-04 | 华北电力大学（保定） | Multitask unloading method and system based on heterogeneous edge cloud architecture |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110531996A (en) * | 2019-08-27 | 2019-12-03 | 河海大学 | Calculating task discharging method based on particle group optimizing under a kind of more thin cloud environment |
| CN111930435A (en) * | 2020-07-13 | 2020-11-13 | 兰州理工大学 | Task unloading decision method based on PD-BPSO technology |
| CN112084025A (en) * | 2020-09-01 | 2020-12-15 | 河海大学 | Improved particle swarm algorithm-based fog calculation task unloading time delay optimization method |
-
2021
- 2021-05-14 CN CN202110532059.5A patent/CN113590211A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110531996A (en) * | 2019-08-27 | 2019-12-03 | 河海大学 | Calculating task discharging method based on particle group optimizing under a kind of more thin cloud environment |
| CN111930435A (en) * | 2020-07-13 | 2020-11-13 | 兰州理工大学 | Task unloading decision method based on PD-BPSO technology |
| CN112084025A (en) * | 2020-09-01 | 2020-12-15 | 河海大学 | Improved particle swarm algorithm-based fog calculation task unloading time delay optimization method |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114143814A (en) * | 2021-12-13 | 2022-03-04 | 华北电力大学（保定） | Multitask unloading method and system based on heterogeneous edge cloud architecture |
| CN114143814B (en) * | 2021-12-13 | 2024-01-23 | 华北电力大学（保定） | Multi-task unloading method and system based on heterogeneous edge cloud architecture |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111586720B (en) | Task unloading and resource allocation combined optimization method in multi-cell scene | |
| WO2022121097A1 (en) | Method for offloading computing task of mobile user | |
| CN111240701B (en) | Task unloading optimization method for end-side-cloud collaborative computing | |
| CN111585816B (en) | Task unloading decision method based on adaptive genetic algorithm | |
| CN112286677B (en) | Resource-constrained edge cloud-oriented Internet of things application optimization deployment method | |
| CN111372314A (en) | Task unloading method and task unloading device based on mobile edge computing scene | |
| CN113242568A (en) | Task unloading and resource allocation method in uncertain network environment | |
| Yoshida et al. | MAB-based client selection for federated learning with uncertain resources in mobile networks | |
| CN112105062B (en) | Mobile edge computing network energy consumption minimization strategy method under time-sensitive condition | |
| CN109246761B (en) | Unloading method based on alternating direction multiplier method considering delay and energy consumption | |
| CN113573363B (en) | MEC calculation unloading and resource allocation method based on deep reinforcement learning | |
| CN113781002B (en) | Low-cost workflow application migration method based on agent model and multiple group optimization in cloud edge cooperative network | |
| CN114650228B (en) | Federal learning scheduling method based on calculation unloading in heterogeneous network | |
| CN113286317A (en) | Task scheduling method based on wireless energy supply edge network | |
| Zhu et al. | Computing offloading strategy using improved genetic algorithm in mobile edge computing system | |
| CN111787543B (en) | 5G communication system resource allocation method based on improved wolf optimization algorithm | |
| CN113992677A (en) | MEC calculation unloading method for delay and energy consumption joint optimization | |
| CN111930435B (en) | Task unloading decision method based on PD-BPSO technology | |
| CN114615730A (en) | Content coverage oriented power distribution method for backhaul limited dense wireless network | |
| CN113590211A (en) | Calculation unloading method based on PSO-DE algorithm | |
| CN114143814A (en) | Multitask unloading method and system based on heterogeneous edge cloud architecture | |
| Baradaran et al. | NEECH: new energy-efficient algorithm based on the best cluster head in wireless sensor networks | |
| CN108965034A (en) | Small-cell base station super-intensive deployment under user-association to network method | |
| CN114615705B (en) | Single-user resource allocation strategy method based on 5G network | |
| CN114980216B (en) | Dependency task unloading system and method based on mobile edge calculation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |