CN112749010B - Edge computing task allocation method of fusion recommendation system - Google Patents
Edge computing task allocation method of fusion recommendation system Download PDFInfo
- Publication number
- CN112749010B CN112749010B CN202011617690.7A CN202011617690A CN112749010B CN 112749010 B CN112749010 B CN 112749010B CN 202011617690 A CN202011617690 A CN 202011617690A CN 112749010 B CN112749010 B CN 112749010B
- Authority
- CN
- China
- Prior art keywords
- edge server
- task
- edge
- cloud
- sender
- 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.)
- Active
Links
- 230000004927 fusion Effects 0.000 title claims abstract description 70
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000004891 communication Methods 0.000 claims description 39
- 238000012545 processing Methods 0.000 claims description 21
- 238000001914 filtration Methods 0.000 claims description 15
- 238000012216 screening Methods 0.000 claims description 11
- 238000005457 optimization Methods 0.000 claims description 6
- 238000005265 energy consumption Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 229920003087 methylethyl cellulose Polymers 0.000 description 1
- 238000013468 resource allocation Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
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
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5061—Partitioning or combining of resources
- G06F9/5072—Grid computing
-
- 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
- G06F9/46—Multiprogramming arrangements
- G06F9/48—Program initiating; Program switching, e.g. by interrupt
- G06F9/4806—Task transfer initiation or dispatching
- G06F9/4843—Task transfer initiation or dispatching by program, e.g. task dispatcher, supervisor, operating system
- G06F9/4881—Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Information Transfer Between Computers (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides an edge computing task allocation method of a fusion recommendation system, which comprises the following steps: step 1, a cloud module, an edge server module and a mobile terminal are mutually connected to construct a cloud-edge-terminal fusion recommendation system; step 2, a task sender sends a task request to a cloud-side-end fusion recommendation system; and 3, the cloud-side-end fusion recommendation system receives a task request sent by a task sender, and the edge server module constructs an edge server database according to the position information and the computing power of a plurality of edge servers which are deployed on the road side facilities and are started to execute tasks according to requirements. The invention combines the cached edge server and the recommended edge server, has high recommended hit rate, proves that the recommended hit rate problem is a monotone sub-module function and NP-hard problem, improves the utilization rate of computer resources, and reduces the time consumption.
Description
Technical Field
The invention relates to the technical field of computer task allocation, in particular to an edge computing task allocation method of a fusion recommendation system.
Background
According to Cisco's report, mobile data traffic will increase 7-fold in the next 5 years, 49 abytes (EB) (1 EB-106 TB) per month will be reached by 2021, while the global IoT device population will increase from 80 to 120 billion at the same time. This makes it difficult for its access network to acquire the computing resources of the mobile edge server.
In addition, in the 5G network era, large-scale task processing demands such as multimedia special effect tasks, big data processing tasks, and the like are also beginning to appear. In these task scenarios, the most common user requirement is real-time requirement, that is, the task is required to be responded and executed quickly, and the execution result is transmitted back to the user quickly. In order to cope with the high-speed development of the mobile internet and the internet of things, the 5G needs to meet the novel business requirements of ultra-low time delay, ultra-low power consumption, ultra-high reliability and ultra-high density connection. Therefore, minimizing task completion time is the most common goal in task offloading issues.
The offloading problems in current mobile edge computing mainly include: and page unloading, namely edge caching, wherein a page provider caches the common pages on an edge cloud so as to reduce time delay and energy consumption when a user requests the pages. There are caching strategies in the related research that take into account page distribution and user mobility. Task offloading, i.e., deciding when, where, how much tasks should be offloaded from the mobile device to the edge to perform, to reduce computation latency and save power consumption. The research is mainly focused on considering the unloading decision problem in a multi-user environment and a multi-server environment. Page offloading is primarily concerned with the storage capacity of the edge cloud, not synchronously considering computing capacity. In the related research of task offloading, it is assumed that the edge cloud has enough software and hardware resources to support task calculation as a normal state, which is contrary to the limitation of the edge cloud resources and the failure to support all types of tasks.
A number of students have conducted related studies with respect to the task offloading problem of MECs. An efficient offloading scheme is proposed by minimizing the total network energy consumption, taking into account the forward and backhaul link capacity constraints and the maximum latency constraints of the users. Under the condition of balancing energy consumption and time delay, an energy-aware calculation unloading scheme is provided, and the residual energy of the battery of the intelligent device is introduced into the definition of the weighting factors of energy consumption and delay, so that the total consumption of the system is effectively reduced. Considering the trade-off between latency and reliability of task offloading, it was studied to divide the tasks of the user equipment into subtasks and offload them in turn to nearby edge nodes. The above documents do not provide for a reasonable allocation of limited wireless and computing resources. In a multi-user MEC system, an online task offloading algorithm is proposed with the goal of minimizing the average energy consumption of the user and the MEC server. Considering the total energy consumption minimization of the system, the joint optimization problem of offloading decisions, radio resource and computing resource allocation is studied. However, existing task scheduling algorithms simply offload and allocate tasks by optimizing energy consumption and delay, but never select an edge server to handle tasks from the task itself.
Disclosure of Invention
The invention provides an edge computing task allocation method of a fusion recommendation system, and aims to solve the problem that an edge server is not selected from tasks per se to process tasks in the traditional task allocation method.
In order to achieve the above objective, an embodiment of the present invention provides a method for assigning an edge computing task of a fusion recommendation system, including:
step 1, a cloud module, an edge server module and a mobile terminal are mutually connected to construct a cloud-edge-terminal fusion recommendation system;
step 2, a task sender sends a task request to a cloud-side-end fusion recommendation system;
step 3, the cloud-side-end fusion recommendation system receives a task request sent by a task sender, and an edge server module constructs an edge server database according to the position information and the computing power of a plurality of edge servers which are deployed on a road side facility and are started to execute tasks according to requirements;
step 4, the cloud-edge-end fusion recommendation system screens out edge server information capable of processing the current task from an edge server database;
step 5, filtering the screened edge server information through a filtering algorithm according to the position information of the current task sender;
step 6, clustering the filtered edge server information according to the computing power of the edge server to obtain the position information and the computing power information of the edge server capable of executing the task request and starting the edge server capable of executing the task request;
step 7, according to the position information and the computing power information of the edge server capable of completing task execution, an edge server table is established in an edge server module, and the edge server module recommends the edge server table to a cloud module;
step 8, the cloud-side-end fusion recommendation system recommends the edge server module to an edge server table of the cloud module to be combined with an edge server which is cached in the cloud module but does not participate in one-stage task execution;
step 9, the cloud-side-end fusion recommendation system screens out edge servers which are not only recommended to the edge servers of the cloud module by the edge server module, but also cached by the cloud module and are not involved in one-stage task execution, and constructs a final recommended edge server information table according to the screened edge servers;
step 10, the cloud-side-end fusion recommendation system analyzes task characteristics of the task request, selects an edge server capable of executing the current task request from a final recommended edge server information table, recommends the edge server to a task sender, and when the task sender receives the edge server recommended by the cloud-side-end fusion recommendation system, the edge server executes the current task request of the task sender;
and 11, optimizing the recommended hit rate of the cloud-side-end fusion recommendation system to the task sender by designing a recommendation optimization algorithm to obtain the optimal recommended hit rate.
Wherein, the step 3 specifically includes:
step 31, extracting and summarizing the deployment position of each edge server, and screening the deployment position information of the edge server;
step 32, summarizing the deployment area of the edge server, and determining the deployment position of the edge server;
step 33, extracting the position of the task sender, screening the deployment positions of the edge servers within the communication range of the task sender, and obtaining an edge server deployment position set within the communication range of the task sender;
step 34, analyzing the computing power of each edge server according to the edge server deployment position set;
step 35, ordering and summarizing the position information of each edge server in the communication range of the task sender, determining the deployment position of each edge server, and establishing a deployment position database of the edge servers;
step 36, summarizing the computing power of each edge server in the communication range of the task sender, and establishing an edge server database participating in task processing.
The step 5 specifically includes:
and step 51, filtering the edge server information in a certain communication range of the current task sender through a filtering algorithm according to the position information of the current task sender transmitted by the edge server module.
Wherein, the step 6 specifically includes:
step 61, separating the edge server information to be recommended from the edge server information after secondary filtering;
and step 62, clustering the edge servers with sufficient computing power when the separated edge server information is excessive, obtaining the position information of the edge servers with sufficient computing power, and starting.
The step 7 specifically includes:
step 71, establishing an edge server table according to the obtained position information of the edge server with sufficient computing power;
in step 72, the edge server module recommends an edge server table to the cloud module.
The step 8 specifically includes:
step 81, the cloud-edge-end fusion recommendation system records an edge server table of the edge server module recommendation cloud module;
and step 82, combining the edge server recommended by the edge server module with the edge server cached by the cloud module.
Wherein, the step 9 and the step 10 specifically include:
and screening out and constructing a final recommended edge server information table for edge servers recommended by the edge server module while being cached in the cloud module, wherein the cloud-edge-end fusion recommendation system uses breadth-first search to add edge servers capable of processing the current task request in the final recommended edge server information table to the tail end of the edge server table, and the cloud-edge-end fusion recommendation system recommends m edge servers capable of completing the current task request to a task sender.
The step 11 specifically includes:
since the task sender is in a mobile state, the edge server is in a flexible state, the task to be processed between the task sender and the edge server changes with time, and the edge server selected to process the task request is in the communication range of the task sender in the current network state:
q ij ＝prob{ED j in the range of TS i}∈{0,1} (1)
wherein q ij Indicating whether the edge server selected to process the task request is within the communication range of the task sender in the current network state, TS i indicating the communication range of the task sender i, EDj indicating the communication range of the edge server j.
Wherein, the step 11 further comprises:
when a task request recommendation of a task sender hits, the recommendation hit rate of the task request is expressed as a function of integer variables, as follows:
where n represents the task request, n.epsilon.K,the probability that the task sender sends the task request n satisfies +.>M represents the number of edge servers, K represents the number of tasks, and x nj Representing the number of edge servers that are used to process task n within the communication range of the task sender.
Wherein, the step 11 further comprises:
the problem of optimizing the recommendation strategy is expressed as:
where N represents the total number of task senders, p ik Representing the probability that task sender i requests to process task k, x nj Represents the number of edge servers that are used to process task n within the communication range of the task sender, and C represents the maximum number of edge servers required to process the task.
The scheme of the invention has the following beneficial effects:
according to the edge computing task allocation method of the fusion recommendation system, which is disclosed by the embodiment of the invention, the cloud-edge-end fusion recommendation system is adopted to perform task feature analysis on a task request of a task sender and select a proper edge server, so that the task is allocated to the proper edge server for execution; the cached edge server is combined with the recommended edge server, the recommended hit rate is high, the problem that the recommended hit rate is NP-hard of a monotonic sub-function is proved, the utilization rate of computer resources is improved, and the time consumption is reduced.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a schematic diagram of the structure of the present invention;
FIG. 3 is a schematic diagram of a recommendation flow chart according to the present invention;
FIG. 4 is a graph showing recommendation rates under different BFS parameters according to the present invention;
FIG. 5 is a schematic diagram of recommendation rates of the present invention in different optimized recommendation modes;
FIG. 6 is a schematic diagram of the time consumption of the present invention under different vehicles;
FIG. 7 is a diagram illustrating the time consumption of the present invention at different task sizes.
Detailed Description
In order to make the technical problems, technical solutions and advantages to be solved more apparent, the following detailed description will be given with reference to the accompanying drawings and specific embodiments.
The invention provides an edge computing task distribution method of a fusion recommendation system, aiming at the problem that an edge server is not selected to process tasks from the task itself in the existing task distribution method.
As shown in fig. 1 to 7, an embodiment of the present invention provides a method for allocating edge computing tasks of a fusion recommendation system, including: step 1, a cloud module, an edge server module and a mobile terminal are mutually connected to construct a cloud-edge-terminal fusion recommendation system; step 2, a task sender sends a task request to a cloud-side-end fusion recommendation system; step 3, the cloud-side-end fusion recommendation system receives a task request sent by a task sender, and an edge server module constructs an edge server database according to the position information and the computing power of a plurality of edge servers which are deployed on a road side facility and are started to execute tasks according to requirements; step 4, the cloud-edge-end fusion recommendation system screens out edge server information capable of processing the current task from an edge server database; step 5, filtering the screened edge server information through a filtering algorithm according to the position information of the current task sender; step 6, clustering the filtered edge server information according to the computing power of the edge server to obtain the position information and the computing power information of the edge server capable of executing the task request and starting the edge server capable of executing the task request; step 7, according to the position information and the computing power information of the edge server capable of completing task execution, an edge server table is established in an edge server module, and the edge server module recommends the edge server table to a cloud module; step 8, the cloud-side-end fusion recommendation system recommends the edge server module to an edge server table of the cloud module to be combined with an edge server which is cached in the cloud module but does not participate in one-stage task execution; step 9, the cloud-side-end fusion recommendation system screens out edge servers which are not only recommended to the edge servers of the cloud module by the edge server module, but also cached by the cloud module and are not involved in one-stage task execution, and constructs a final recommended edge server information table according to the screened edge servers; step 10, the cloud-side-end fusion recommendation system analyzes task characteristics of the task request, selects an edge server capable of executing the current task request from a final recommended edge server information table, recommends the edge server to a task sender, and when the task sender receives the edge server recommended by the cloud-side-end fusion recommendation system, the edge server executes the current task request of the task sender; and 11, optimizing the recommended hit rate of the cloud-side-end fusion recommendation system to the task sender by designing a recommendation optimization algorithm to obtain the optimal recommended hit rate.
In the method for assigning an edge computing task of a fusion recommendation system according to the above embodiment of the present invention, assuming that an edge server information record is regarded as a series of data points, each edge server ID-computing capability-location data point is: s= (id, calculating power, position); all edge server information can be noted as: q= (S 1 ,S 2 ,…,S n ) The method comprises the steps of carrying out a first treatment on the surface of the Since in practice, the fixed deployment location of the edge server is only started when needed, it is necessary to roughly define the edge servers within a certain communication range based on the location of the task sender, and filter out the location information of the edge servers. The obtaining of the edge server position information within the specified range is divided into: extracting an edge server deployment position, generalizing an edge server deployment area and analyzing the computing capacity of the edge server. 1. Extracting edge server deployment locations includes extracting and generalizing individual edge serversAnd (5) the deployment position and the edge server deployment position information are screened. 2. Summarizing the edge server deployment area includes determining, screening, merging edge server deployment locations, and summarizing the edge server deployment area. 3. Analyzing the edge server computing capabilities includes ordering edge server deployment areas, screening edge server deployment areas, analyzing edge server computing capabilities. Each deployment location indicates that the edge server has occurred at a certain location with a certain meaning. The method comprises the steps of extracting and summarizing deployment positions to obtain an edge server deployment position set in a communication range of a task sender, analyzing the computing capacity condition of each edge server, finally sequencing and summarizing the edge servers, determining the positions of the edge servers, establishing an edge server position database, summarizing the computing capacity of the edge servers in a fixed position area, and establishing the edge server database.
Wherein, the step 3 specifically includes: step 31, extracting and summarizing the deployment position of each edge server, and screening the deployment position information of the edge server; step 32, summarizing the deployment area of the edge server, and determining the deployment position of the edge server; step 33, extracting the position of the task sender, screening the deployment positions of the edge servers within the communication range of the task sender, and obtaining an edge server deployment position set within the communication range of the task sender; step 34, analyzing the computing power of each edge server according to the edge server deployment position set; step 35, ordering and summarizing the position information of each edge server in the communication range of the task sender, determining the deployment position of each edge server, and establishing a deployment position database of the edge servers; step 36, summarizing the computing power of each edge server in the communication range of the task sender, and establishing an edge server database participating in task processing.
The step 5 specifically includes: and step 51, filtering the edge server information in a certain communication range of the current task sender through a filtering algorithm according to the position information of the current task sender transmitted by the edge server module.
Wherein, the step 6 specifically includes: step 61, separating the edge server information to be recommended from the edge server information after secondary filtering; and step 62, clustering the edge servers with sufficient computing power when the separated edge server information is excessive, obtaining the position information of the edge servers with sufficient computing power, and starting.
The step 7 specifically includes: step 71, establishing an edge server table according to the obtained position information of the edge server with sufficient computing power; in step 72, the edge server module recommends an edge server table to the cloud module.
The step 8 specifically includes: step 81, the cloud-edge-end fusion recommendation system records an edge server table of the edge server module recommendation cloud module; and step 82, combining the edge server recommended by the edge server module with the edge server cached by the cloud module.
Wherein, the step 9 and the step 10 specifically include: and screening out and constructing a final recommended edge server information table for edge servers recommended by the edge server module while being cached in the cloud module, wherein the cloud-edge-end fusion recommendation system uses breadth-first search to add edge servers capable of processing the current task request in the final recommended edge server information table to the tail end of the edge server table, and the cloud-edge-end fusion recommendation system recommends m edge servers capable of completing the current task request to a task sender.
According to the edge computing task allocation method of the fusion recommendation system, the edge server information table contains recommended edge server information and cloud-cached edge server information, and the cloud-side-end fusion recommendation system selects a proper edge server from the edge server information table based on the content relation of the task request and recommends the proper edge server to a task sender.
The step 11 specifically includes: since the task sender is in a mobile state, the edge server is in a flexible state, the task to be processed between the task sender and the edge server changes with time, and the edge server selected to process the task request is in the communication range of the task sender in the current network state:
q ij ＝prob{ED j in the range of TS i}∈{0,1} (1)
wherein q ij Indicating whether the edge server selected to process the task request is within the communication range of the task sender in the current network state, TS i indicating the communication range of the task sender i, EDj indicating the communication range of the edge server j.
Wherein, the step 11 further comprises: when a task request recommendation of a task sender hits, the recommendation hit rate of the task request is expressed as a function of integer variables, as follows:
where n represents the task request, n.epsilon.K,the probability that the task sender sends the task request n satisfies +.>M represents the number of edge servers, K represents the number of tasks, and x nj Representing the number of edge servers that are used to process task n within the communication range of the task sender.
Wherein, the step 11 further comprises: the problem of optimizing the recommendation strategy is expressed as:
where N represents the total number of task senders, p ik Representing the probability that task sender i requests to process task k, x nj Represents the number of edge servers that are used to process task n within the communication range of the task sender, and C represents the maximum number of edge servers required to process the task.
According to the edge computing task allocation method of the fusion recommendation system, a task sender sends a task request to the cloud-edge-end fusion recommendation system, edge server information capable of completing tasks is searched according to the following BFS mode, first, the edge server information related to task content is requested to the cloud-edge-end fusion recommendation system, and the sequence of the requested edge server information returned from the cloud-edge-end fusion recommendation system is added to the list M. Edge server information associated therewith and capable of task processing is further recommended for edge server information in list M and added to the end of list M. And the like, adding other relevant edge server information until the search depth is reached, searching the list M for the edge server information simultaneously contained in the cloud cache through a TORS algorithm, adding the edge server information into an output list G until all the edge server information in the list M is browsed, and outputting the list G to contain N edge server information, wherein the list M is the first to come (line 5-10). When the above steps are completed, the number of edge servers in the output list G is less than N, and N-G| pieces of edge server information are added to the output list G (lines 11-16).
According to the edge computing task allocation method of the fusion recommendation system, disclosed by the embodiment of the invention, the optimization problem of the formula (3) is proved as follows:
the objective function of the optimization problem is NP-hard problem, usingRepresenting a set of edge servers associated with a current processing task k, as follows:
wherein t represents an edge server different from the edge server j for processing the current task k, and M represents the number of edge servers;
the edge server subset is represented asSuppose that only edge server j is recommended to handle the task (x j ＝1 and x t ＝0,/>) Therefore the recommendation rate will be equal to +.>(/>Is the probability that the edge server t processes task i). When not only edge server j is recommended to handle a task, S' represents the union of all edge servers covered by the current task +.>Therefore the recommendation rate is equal to +.>The objective function thus corresponds to:
wherein,representing the probability of the edge server t processing task i, q it Indicating the sender communication range of the edge server t at task i.
This corresponds to the maximum coverage problem of weighted elements, where elements correspond to edge servers, weights correspond to probability values, and the number of subsets selected must be less than the maximum number of edge servers required for processing tasks, and the covered elements are combined to S'. This is a known NP-hard problem. Thus, there are many edge server problems recommended within the scope of each task coverage, which is also an NP-hard problem.
The objective function is a monotonic sub-model function and is radix-constrained: objective function f (u):equivalent to the aggregate function f (S)>Where KxM is the edge server recommendation set (K, j), K represents the task, j represents the edge server, S= { K ε K, i ε M: x kj =1 }, as follows:
wherein p is ik Representing the probability that task sender i requests processing task k,the probability that the task sender sends the task request n satisfies +.>q ij The communication range of the task sender is represented by i, the task sender is represented by j, and the edge server is represented by j.
One set function is characterized as a sub-set function if and only if, for eachAnd ε V\B is considered:
[f(A∪{ε})-f(A)]-[f(E∪{ε})-f(E)]≥0 (7)
wherein A and E represent sets, satisfy for all A and E setsEpsilon represents an element belonging to V and not belonging to E;
according to the formulaFirst, calculating:
wherein l, ε represents an element other than A, p ik Representing the probability that task sender i requests processing task k,the probability that the task sender sends the task request n satisfies the requirement when the task k is unreasonableq ij And q im The communication range of the task sender is represented, i represents the task sender, j and m represents the edge server.
The value of formula (9) is always greater than 0, i.e. the sub-modularity is demonstrated;
monotonicity is demonstrated because equation (10) is always non-negative;
to prove that the constraint is a pseudo-array, consider the set v=k×m (i.e., all possible tuples { task, edge servers }) and the set of subsets thereof without violating the maximum number of edge servers required to process the task, as follows:
wherein S represents a subset of edge servers, 2 ν Representing the power set of the set v, C represents the edge services required for processing tasksMaximum number of servers, M represents the number of edge servers.
First, for all sets A and E,it thinks that if->The number of edge servers recommended by set E does not exceed the maximum number of edge servers required for task processing, then for +.>Because each edge server in set a must be the same as or less than the edge servers in set E, the edge server number problem is not violated. Secondly, for all collections A, E Γ, edge servers meeting recommended conditions, |E|l>More edge servers are recommended in set E, in which the number of edge servers does not reach a maximum, otherwise set E would violate the edge server number constraint, i.e. & lt/EN & gt>This means that set a may continue to recommend at least one or more edge servers, and that edge server may be selected from set E;
for any monotonic sub-module function f, it is considered that:
F(x)≥(1-1/e)f * (x) (12)
wherein f * (x) Representing the optimal value of the monotonic sub-function.
In order to maximize the monotonic sub-functions constrained by the pseudo-arrays, khiller S proposes a stochastic algorithm that gives a (1-1/e) approximation, which is split into two parts. In the first part, the combined problem is replaced with a continuous problem and an approximate solution to the continuous problem is found. In the second section, the fractional solution of the continuous problem is rounded using a technique called Pipage round. Although this algorithm provides better performance guarantees, it is computationally too computationally intensive to implement when the edge server recommendation set size in question is equal to k×m. Of course, the minimization algorithm complexity or best approximation algorithm is beyond the scope of what is needed to be processed. A fast and effective Greedy algorithm Greedy is used for processing a big task in a cloud edge cooperative scene.
In the method for allocating the edge computing tasks of the fusion recommendation system according to the above embodiment of the present invention, fig. 4 shows that by directly recommending and caching, the recommendation rate of recommending and caching is higher than that of directly recommending as the parameters B and D increase. The recommendation rate performance of TORS algorithm, greedy algorithm, and Top recommendation algorithm under different BFS parameters is shown in fig. 5. For the same case, greedy algorithm Greedy is always better than TORS algorithm, and the recommendation rate performance of Top recommendation algorithm is 2 times higher than Greedy algorithm Greedy, and the benefit of combining recommendation and caching is clearly shown in fig. 5.
The method for allocating edge computing tasks of the fusion recommendation system according to the above embodiment of the present invention, in fig. 6 and fig. 7, illustrates the total cost of time of the system when the number of edge nodes increases. In general, the total cost of the three approaches decreases as the number of edge nodes increases. In fig. 6, the Top recommendation algorithm is optimal, the Greedy algorithm Greedy is very small in difference, and both the TORS algorithm and Greedy algorithm Greedy are relatively stable. Meanwhile, the Top recommendation algorithm optimal recommendation curve is higher than the Greedy algorithm Greedy optimal recommendation curve, and the single-side nodes are Greedy. But the more edge nodes, the faster they drop. This is because the execution time decreases as the number of edge nodes increases. Fig. 7 shows the overall cost impact of off-load task data size on the system, where the number of edge nodes is set to 4, with the overall cost of the three approaches increasing as the off-load task data size increases. This is because the larger the data volume, the greater the time and energy consumption of the offload. Compared with other methods, the TORS algorithm has slower growth trend and better effect. With the increase of the data volume, the increase speed of the Top recommendation algorithm curve is far faster than that of the TORS algorithm and Greedy algorithm Greedy, which shows that the larger the data volume required for unloading is, the larger the delay and energy consumption of the unloading calculation are.
According to the edge computing task allocation method of the fusion recommendation system, which is disclosed by the embodiment of the invention, the cloud-edge-end fusion recommendation system is adopted to perform task feature analysis on the task request of the task sender and select a proper edge server, so that the task is allocated to the proper edge server for execution. The cloud-side-end fusion recommendation system mainly comprises three modules: cloud module, edge server module and mobile terminal (task sender). The edge server module mainly recommends an edge server in a flexible state (not cached), and the edge server meeting the condition is sufficiently recommended to the cloud module according to whether the edge server is in communication range and computing power of a task sender. The cloud module combines the edge server information acquired from the edge server module with the edge server cached in the cloud, and recommends the information to the task sender. The cloud-side-end fusion recommendation system is designed by combining the cached edge servers with the recommended edge servers, has higher recommendation hit rate, and proves that the problem of the recommendation hit rate is a monotone sub-module function and an NP-hard problem, so that the utilization rate of computer resources of the whole system is improved, and the time consumption is reduced.
While the foregoing is directed to the preferred embodiments of the present invention, it will be appreciated by those skilled in the art that various modifications and adaptations can be made without departing from the principles of the present invention, and such modifications and adaptations are intended to be comprehended within the scope of the present invention.
Claims (7)
1. The edge computing task allocation method of the fusion recommendation system is characterized by comprising the following steps of:
step 1, a cloud module, an edge server module and a mobile terminal are mutually connected to construct a cloud-edge-terminal fusion recommendation system;
step 2, a task sender sends a task request to a cloud-side-end fusion recommendation system;
step 3, the cloud-side-end fusion recommendation system receives a task request sent by a task sender, and an edge server module constructs an edge server database according to the position information and the computing power of a plurality of edge servers which are deployed on a road side facility and are started to execute tasks according to requirements;
step 4, the cloud-edge-end fusion recommendation system screens out edge server information capable of processing the current task from an edge server database;
step 5, filtering the screened edge server information through a filtering algorithm according to the position information of the current task sender;
step 6, clustering the filtered edge server information according to the computing power of the edge server to obtain the position information and the computing power information of the edge server capable of executing the task request and starting the edge server capable of executing the task request;
step 7, according to the position information and the computing power information of the edge server capable of completing task execution, an edge server table is established in an edge server module, and the edge server module recommends the edge server table to a cloud module;
step 8, the cloud-side-end fusion recommendation system recommends the edge server module to an edge server table of the cloud module to be combined with an edge server which is cached in the cloud module but does not participate in one-stage task execution;
step 9, the cloud-side-end fusion recommendation system screens out edge servers which are not only recommended to the edge servers of the cloud module by the edge server module, but also cached by the cloud module and are not involved in one-stage task execution, and constructs a final recommended edge server information table according to the screened edge servers;
step 10, the cloud-side-end fusion recommendation system analyzes task characteristics of the task request, selects an edge server capable of executing the current task request from a final recommended edge server information table, recommends the edge server to a task sender, and when the task sender receives the edge server recommended by the cloud-side-end fusion recommendation system, the edge server executes the current task request of the task sender;
step 11, optimizing the recommended hit rate of the cloud-side-end fusion recommendation system to the task sender by designing a recommendation optimization algorithm to obtain an optimal recommended hit rate; the step 11 specifically includes:
since the task sender is in a mobile state, the edge server is in a flexible state, the task to be processed between the task sender and the edge server changes with time, and the edge server selected to process the task request is in the communication range of the task sender in the current network state:
q ij ＝prob{ED j in the range of TS i}∈{0,1} (1)
wherein q ij Indicating whether the edge server selected to process the task request is within the communication range of the task sender in the current network state, i indicating the task sender, TSi indicating the communication range of the task sender i, EDj indicating the communication range of the edge server j; the step 11 further includes:
when a task request recommendation of a task sender hits, the recommendation hit rate of the task request is expressed as a function of integer variables, as follows:
where n represents the task request, n.epsilon.K,the probability that the task sender sends the task request n satisfies +.>M represents the number of edge servers, K represents the number of tasks, and x nj Representing the number of edge servers used to process task n within communication range of the task sender; the step 11 further includes:
the problem of optimizing the recommendation strategy is expressed as:
where N represents the total of task sendersQuantity, p ik Representing the probability that task sender i requests to process task k, x nj Represents the number of edge servers that are used to process task n within the communication range of the task sender, and C represents the maximum number of edge servers required to process the task.
2. The method for assigning edge computing tasks of a fusion recommendation system according to claim 1, wherein the step 3 specifically comprises:
step 31, extracting and summarizing the deployment position of each edge server, and screening the deployment position information of the edge server;
step 32, summarizing the deployment area of the edge server, and determining the deployment position of the edge server;
step 33, extracting the position of the task sender, screening the deployment positions of the edge servers within the communication range of the task sender, and obtaining an edge server deployment position set within the communication range of the task sender;
step 34, analyzing the computing power of each edge server according to the edge server deployment position set;
step 35, ordering and summarizing the position information of each edge server in the communication range of the task sender, determining the deployment position of each edge server, and establishing a deployment position database of the edge servers;
step 36, summarizing the computing power of each edge server in the communication range of the task sender, and establishing an edge server database participating in task processing.
3. The method for assigning edge computing tasks of a fusion recommendation system according to claim 1, wherein the step 5 specifically comprises:
and step 51, filtering the edge server information in a certain communication range of the current task sender through a filtering algorithm according to the position information of the current task sender transmitted by the edge server module.
4. The method for assigning edge computing tasks of a fusion recommendation system according to claim 1, wherein the step 6 specifically comprises:
step 61, separating the edge server information to be recommended from the edge server information after secondary filtering;
and step 62, clustering the edge servers with sufficient computing power when the separated edge server information is excessive, obtaining the position information of the edge servers with sufficient computing power, and starting.
5. The method for assigning edge computing tasks of a fusion recommendation system according to claim 1, wherein the step 7 specifically comprises:
step 71, establishing an edge server table according to the obtained position information of the edge server with sufficient computing power;
in step 72, the edge server module recommends an edge server table to the cloud module.
6. The method for assigning edge computing tasks of a fusion recommendation system according to claim 1, wherein the step 8 specifically comprises:
step 81, the cloud-edge-end fusion recommendation system records an edge server table of the edge server module recommendation cloud module;
and step 82, combining the edge server recommended by the edge server module with the edge server cached by the cloud module.
7. The method for assigning an edge computing task to a fusion recommendation system according to claim 1, wherein the step 9 and the step 10 specifically comprise:
and screening out and constructing a final recommended edge server information table for edge servers recommended by the edge server module while being cached in the cloud module, wherein the cloud-edge-end fusion recommendation system uses breadth-first search to add edge servers capable of processing the current task request in the final recommended edge server information table to the tail end of the edge server table, and the cloud-edge-end fusion recommendation system recommends m edge servers capable of completing the current task request to a task sender.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011617690.7A CN112749010B (en) | 2020-12-31 | 2020-12-31 | Edge computing task allocation method of fusion recommendation system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011617690.7A CN112749010B (en) | 2020-12-31 | 2020-12-31 | Edge computing task allocation method of fusion recommendation system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112749010A CN112749010A (en) | 2021-05-04 |
| CN112749010B true CN112749010B (en) | 2024-03-12 |
Family
ID=75650148
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011617690.7A Active CN112749010B (en) | 2020-12-31 | 2020-12-31 | Edge computing task allocation method of fusion recommendation system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112749010B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113242319A (en) * | 2021-07-09 | 2021-08-10 | 军事科学院系统工程研究院网络信息研究所 | Cloud edge cooperative work method and system based on video cloud service architecture |
| CN113626107B (en) * | 2021-08-20 | 2024-03-26 | 中南大学 | Mobile computing unloading method, system and storage medium |
| CN113722531A (en) * | 2021-09-08 | 2021-11-30 | 北京奇艺世纪科技有限公司 | Data storage method, device, system, equipment and storage medium |
| CN113572857B (en) * | 2021-09-23 | 2021-12-03 | 环球数科集团有限公司 | Resource management system applied to Internet of things edge cloud collaborative computing environment |
| CN113919512B (en) * | 2021-09-26 | 2022-09-23 | 重庆邮电大学 | Federal learning communication optimization method and system based on computing resource logic layering |
| CN114327914B (en) * | 2022-03-07 | 2022-05-06 | 中国气象局公共气象服务中心(国家预警信息发布中心) | Mountain scenic spot mountain climbing decision-making method and system based on multi-factor edge calculation |
| CN117971503B (en) * | 2024-03-29 | 2024-06-11 | 杭州远石科技有限公司 | Data caching method and system based on edge calculation |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107612987A (en) * | 2017-09-08 | 2018-01-19 | 浙江大学 | A kind of service provision optimization method based on caching towards edge calculations |
| CN109347924A (en) * | 2018-09-20 | 2019-02-15 | 西北大学 | A kind of recommended method based on intelligent perception |
| CN110213627A (en) * | 2019-04-23 | 2019-09-06 | 武汉理工大学 | Flow medium buffer distributor and its working method based on multiple cell user mobility |
| CN111556531A (en) * | 2020-04-28 | 2020-08-18 | 东南大学 | Cooperative cache optimization method in micro-cellular wireless network |
-
2020
- 2020-12-31 CN CN202011617690.7A patent/CN112749010B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107612987A (en) * | 2017-09-08 | 2018-01-19 | 浙江大学 | A kind of service provision optimization method based on caching towards edge calculations |
| CN109347924A (en) * | 2018-09-20 | 2019-02-15 | 西北大学 | A kind of recommended method based on intelligent perception |
| CN110213627A (en) * | 2019-04-23 | 2019-09-06 | 武汉理工大学 | Flow medium buffer distributor and its working method based on multiple cell user mobility |
| CN111556531A (en) * | 2020-04-28 | 2020-08-18 | 东南大学 | Cooperative cache optimization method in micro-cellular wireless network |
Non-Patent Citations (9)
| Title |
|---|
| "Computation Placement Orchestrator for Mobile Edge Computing in Heterogeneous Vehicular Networks";L.Wang, X. Deng, J. Gui, H. Zhang and S. Yu;《in IEEE Internet of Things Journal》;全文 * |
| "Joint Optimization of Caching and Computation in Multi-Server NOMA-MEC System via Reinforcement Learning";S.Li, B. Li and W. Zhao;《in IEEE Access》;全文 * |
| "Pccp: Proactive video chunks caching and processing in edge networks.";Baccour, Emna, et al.;《Future Generation Computer Systems》;全文 * |
| "Proactive content caching by exploiting transfer learning for mobile edge computing.";Hou, Tingting, et al.;《International Journal of Communication Systems》;全文 * |
| "Smart caching based on user behavior for mobile edge computing.";Zeng, Yuanyuan, et al.;《Information Sciences》;全文 * |
| "CABaRet: Leveraging recommendation systems for mobile edge caching";Kastanakis S, Sermpezis P, Kotronis V, et al.;《Proceedings of the 2018 Workshop on Mobile Edge Communications》;全文 * |
| "移动边缘计算（MEC）架构中高效内容分发技术";侯停停;《中国优秀硕士学位论文全文数据库 信息科技辑》;全文 * |
| 乐光学 ; 戴亚盛 ; 杨晓慧 ; 刘建华 ; 游真旭 ; 朱友康."边缘计算可信协同服务策略建模".《计算机研究与发展》.2020,全文. * |
| 邓晓衡,关培源,万志文等.."基于综合信任的边缘计算资源协同研究".《计算机研究与发展》.2018,全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112749010A (en) | 2021-05-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112749010B (en) | Edge computing task allocation method of fusion recommendation system | |
| CN111836283B (en) | Internet of vehicles resource allocation method based on MEC multi-server | |
| CN112020103B (en) | Content cache deployment method in mobile edge cloud | |
| CN111010684B (en) | Internet of vehicles resource allocation method based on MEC cache service | |
| CN110087257B (en) | Task unloading device and method supporting mobile edge calculation | |
| CN111031102A (en) | Multi-user, multi-task mobile edge computing system cacheable task migration method | |
| CN108495340B (en) | Network resource allocation method and device based on heterogeneous hybrid cache | |
| CN111475274B (en) | Cloud collaborative multi-task scheduling method and device | |
| CN113918240B (en) | Task unloading method and device | |
| He et al. | QoE-based cooperative task offloading with deep reinforcement learning in mobile edge networks | |
| Huang et al. | Delay-aware caching in internet-of-vehicles networks | |
| CN112540845B (en) | Collaboration system and method based on mobile edge calculation | |
| CN106792995B (en) | User access method for guaranteeing low-delay content transmission in 5G network | |
| CN114007225A (en) | BWP allocation method, apparatus, electronic device and computer readable storage medium | |
| Cha et al. | Fuzzy logic based client selection for federated learning in vehicular networks | |
| CN112235385A (en) | Offshore edge computing low-overhead cloud edge intelligent cooperative ally member discovery method | |
| CN115714814B (en) | Edge cache replacement method based on multi-agent reinforcement learning | |
| CN114641041A (en) | Edge-intelligent-oriented Internet of vehicles slicing method and device | |
| CN115967990A (en) | Classification and prediction-based border collaborative service unloading method | |
| CN113992677A (en) | MEC calculation unloading method for delay and energy consumption joint optimization | |
| Hao et al. | Energy-aware offloading based on priority in mobile cloud computing | |
| CN111654873B (en) | Mobile CDN link selection energy consumption optimization method based on global utility cache strategy | |
| Li et al. | Video caching and scheduling with edge cooperation | |
| Fang et al. | Offloading strategy for edge computing tasks based on cache mechanism | |
| CN113766540B (en) | Low-delay network content transmission method, device, electronic equipment and medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |