CN116634500A - D2D computing unloading method based on hypergraph matching computation and communication capacity enhancement - Google Patents

D2D computing unloading method based on hypergraph matching computation and communication capacity enhancement Download PDF

Info

Publication number
CN116634500A
CN116634500A CN202310628662.2A CN202310628662A CN116634500A CN 116634500 A CN116634500 A CN 116634500A CN 202310628662 A CN202310628662 A CN 202310628662A CN 116634500 A CN116634500 A CN 116634500A
Authority
CN
China
Prior art keywords
optimization model
representing
task
offload
constraint
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
Application number
CN202310628662.2A
Other languages
Chinese (zh)
Inventor
赵攀
蒋志良
吴绍冲
刘保菊
徐大同
崔名扬
李达
黄星硕
谢占龙
贾翔源
陈鑫鑫
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Henan University of Technology
Original Assignee
Henan University of Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Henan University of Technology filed Critical Henan University of Technology
Priority to CN202310628662.2A priority Critical patent/CN116634500A/en
Publication of CN116634500A publication Critical patent/CN116634500A/en
Pending legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W28/00—Network traffic management; Network resource management
    • H04W28/02—Traffic management, e.g. flow control or congestion control
    • H04W28/08—Load balancing or load distribution
    • H04W28/0875—Load balancing or load distribution to or through Device to Device [D2D] links, e.g. direct-mode links
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W28/00—Network traffic management; Network resource management
    • H04W28/02—Traffic management, e.g. flow control or congestion control
    • H04W28/08—Load balancing or load distribution
    • H04W28/09—Management thereof
    • H04W28/0925—Management thereof using policies
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W28/00—Network traffic management; Network resource management
    • H04W28/02—Traffic management, e.g. flow control or congestion control
    • H04W28/08—Load balancing or load distribution
    • H04W28/09—Management thereof
    • H04W28/0958—Management thereof based on metrics or performance parameters
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W28/00—Network traffic management; Network resource management
    • H04W28/02—Traffic management, e.g. flow control or congestion control
    • H04W28/08—Load balancing or load distribution
    • H04W28/09—Management thereof
    • H04W28/0958—Management thereof based on metrics or performance parameters
    • H04W28/0967—Quality of Service [QoS] parameters
    • H04W28/0975—Quality of Service [QoS] parameters for reducing delays
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W28/00—Network traffic management; Network resource management
    • H04W28/02—Traffic management, e.g. flow control or congestion control
    • H04W28/08—Load balancing or load distribution
    • H04W28/09—Management thereof
    • H04W28/0958—Management thereof based on metrics or performance parameters
    • H04W28/0967—Quality of Service [QoS] parameters
    • H04W28/0983—Quality of Service [QoS] parameters for optimizing bandwidth or throughput
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W52/00—Power management, e.g. TPC [Transmission Power Control], power saving or power classes
    • H04W52/04—TPC
    • H04W52/18—TPC being performed according to specific parameters
    • H04W52/26—TPC being performed according to specific parameters using transmission rate or quality of service QoS [Quality of Service]
    • H04W52/265—TPC being performed according to specific parameters using transmission rate or quality of service QoS [Quality of Service] taking into account the quality of service QoS
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W52/00—Power management, e.g. TPC [Transmission Power Control], power saving or power classes
    • H04W52/04—TPC
    • H04W52/18—TPC being performed according to specific parameters
    • H04W52/26—TPC being performed according to specific parameters using transmission rate or quality of service QoS [Quality of Service]
    • H04W52/267—TPC being performed according to specific parameters using transmission rate or quality of service QoS [Quality of Service] taking into account the information rate
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W72/00—Local resource management
    • H04W72/04—Wireless resource allocation
    • H04W72/044—Wireless resource allocation based on the type of the allocated resource
    • H04W72/0453—Resources in frequency domain, e.g. a carrier in FDMA
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W72/00—Local resource management
    • H04W72/04—Wireless resource allocation
    • H04W72/044—Wireless resource allocation based on the type of the allocated resource
    • H04W72/0473—Wireless resource allocation based on the type of the allocated resource the resource being transmission power
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W72/00—Local resource management
    • H04W72/20—Control channels or signalling for resource management
    • H04W72/25—Control channels or signalling for resource management between terminals via a wireless link, e.g. sidelink
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W72/00—Local resource management
    • H04W72/50—Allocation or scheduling criteria for wireless resources
    • H04W72/535—Allocation or scheduling criteria for wireless resources based on resource usage policies
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W72/00—Local resource management
    • H04W72/50—Allocation or scheduling criteria for wireless resources
    • H04W72/54—Allocation or scheduling criteria for wireless resources based on quality criteria
    • H04W72/542—Allocation or scheduling criteria for wireless resources based on quality criteria using measured or perceived quality
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W72/00—Local resource management
    • H04W72/50—Allocation or scheduling criteria for wireless resources
    • H04W72/54—Allocation or scheduling criteria for wireless resources based on quality criteria
    • H04W72/543—Allocation or scheduling criteria for wireless resources based on quality criteria based on requested quality, e.g. QoS
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Quality & Reliability (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The D2D computation unloading party based on hypergraph matching computation and communication capacity enhancement relates to the technical field of model production, and comprises the following steps: acquiring channel state information and task state information to construct a first network optimization model; performing optimal decoupling on the data processing result of the first network optimization model to generate a second optimization model based on an unloading strategy and channel allocation, and a third optimization model based on power control and task allocation; performing data analysis of unloading strategy and channel allocation in a second optimization model based on hypergraph matching, conflict graphs, independent set principles and second constraint conditions, and outputting an optimization result; in the third optimization model, based on the optimization result output by the second optimization model, carrying out data analysis by using an iterative search algorithm and a third constraint condition, and then outputting D2D calculation unloading result data under spectrum sharing; the task execution cost of the unloading method executed by the application is obviously reduced, and the frequency spectrum efficiency and the calculation efficiency of the network are improved.

Description

D2D computing unloading method based on hypergraph matching computation and communication capacity enhancement
Technical Field
The application relates to the technical field of model production, in particular to a D2D computing unloading method based on hypergraph matching computing and communication capacity enhancement.
Background
With the development of 5G business and 6G, the number of internet of things (Internet of Things, ioT) devices and traffic has increased exponentially, and it is predicted that the number of IoT devices will reach 270 billions in 2025 years, and at the same time, the internet of things devices are widely applied to smart cities, households, autopilots, mobile medical treatment, and other scenes; however, the intelligent development of these applications faces challenges in both computing and communication; 1) In terms of computing, resource-constrained internet of things devices are difficult to meet the requirements of novel applications for computing processing capacity; 2) In the aspect of communication, the limited network spectrum resource can not guarantee the transmission requirements of large-scale Internet of things connection and massive calculation data; therefore, how to realize dynamic regulation and flexible allocation of resources under the conditions of relatively fixed resources and cost control is required, so as to meet the dual requirements of the Internet of things on calculation and communication, and the method is a core problem of the current Internet of things development;
the traditional mobile edge computing (Mobile Edge Computing, MEC) technology makes up the shortages of the Internet of things equipment in terms of computing capacity and the like by migrating complex computing tasks to an edge server with stronger computing capacity; however, the expensive installation and maintenance costs make the MEC server have limited computational resources, and the large-scale computational demands will cause the problem of degradation of quality of service due to resource competition; in addition, mobile edge computing offloading performance is highly dependent on wireless access efficiency, facing large-scale links and massive computing data, limited spectrum resources and poor cellular connectivity are obstacles for their further development; meanwhile, a large number of mobile terminal devices exist in a network, and the computing resources of most mobile devices are in an idle state, so D2D (Devices To Devices) computing and unloading are proposed, and the computing tasks are transmitted to the neighboring devices with idle computing resources for processing through D2D links with higher transmission efficiency by utilizing the intensity heterogeneity and the run-time asynchronism of the available computing resources among the mobile devices; D2D computational offloading is particularly attractive for emerging 5G networks due to its proximity, low latency, better coverage and traffic offload gain characteristics; in order to further improve computing and communication capabilities of the edge network, and computing offload performance, a D2D computing offload architecture has attracted extensive attention;
notably, most existing work in studying D2D offload scenarios assumes that D2D computation offloaded data communications employ dedicated spectrum; however, there are some fundamental challenges to be addressed in minimizing the computational offload overhead costs in shared spectrum; in particular, spectral multiplexing can cause cross-layer interference, thereby increasing transmission delay; in the case of communication resource allocation without interference perception, the computation delay may be deteriorated, and the device transmit power may be wasted; in addition, considering the execution mode of tasks, user association and strong coupling between resource sharing, resource allocation is more challenging, and particularly, the user association and channel resource sharing bring about the compromise between offloading benefit and communication overhead; therefore, considering the limitation of computing and communication resources in combination, combining user association, channel resource sharing and power control, designing a D2D offloading mechanism for reducing the total overhead of system offloading becomes a focus of attention, and how to balance the offloading benefit and the communication cost to perform the optimal pairing and resource allocation scheme of offloading devices in the presence of co-channel interference is a technical problem to be solved.
Disclosure of Invention
In order to overcome the defects in the background art, the application discloses a D2D calculation unloading method based on hypergraph matching calculation and communication capacity enhancement, which is characterized in that a mixed integer nonlinear programming (MINLP) optimization model for minimizing task execution overhead is established, constraints such as local energy consumption, task completion expiration date, maximum power and the like are comprehensively considered, a system optimization model for combining task segmentation, power distribution, user association and channel distribution is formulated, and the system optimization model is designed to be based on a first model for unloading strategy and channel distribution and a second model based on power control and task distribution, and continuous optimization problems of the power control and the task segmentation are solved through a search method, so that the maximization of the system calculation and the communication capacity is realized.
In order to achieve the aim of the application, the application adopts the following technical scheme: a hypergraph matching calculation and communication capacity enhancement-based D2D calculation offloading method, comprising the steps of: s1, obtaining channel state information and task state information to construct a first network optimization model, and setting a first constraint condition of the first network optimization model; s2, performing optimal decoupling on the data processing result of the first network optimization model, generating a second optimization model based on an unloading strategy and channel allocation and a third optimization model based on power control and task allocation, and setting a second constraint condition and a third constraint condition based on the second optimization model and the third optimization model respectively; s3, carrying out data analysis of unloading strategy and channel allocation in a second optimization model based on hypergraph matching, conflict graphs, independent set principles and second constraint conditions, and outputting an optimization result; s4, in the third optimization model, based on the optimization result output by the second optimization model, performing data analysis by using an iterative search algorithm and a third constraint condition, and outputting D2D calculation unloading result data.
Further, in S1, a user equipment data set based on a first network optimization model is constructed, including a task user equipment TD, a service user equipment SD with idle resources, and a cellular user equipment CD, where m= {1,2,..m } represents a set of TDs, M is the number of TDs, and mth TD is represented as TDm; SD is a set of resource-free user equipments with high computational power; let s= {1,2,..s } denote a set of SDs, where S is the number of SDs, and S-th SD is denoted as SDs; the CD is a cellular user that has communicated with the base station, and the channel thereof is multiplexed by D2D offload data transmission, where n= {1,2,..n } represents a set of CDs, where N is the number of CDs, and the nth CD is denoted as CDn;
generating a first network optimization model, expressed by the following formula:
in the formula ,δms Representing a user association relationship between a task execution mode of the user TD and an unloading strategy; q (Q) m Representing the overall cost, including power consumption and time delay,representing channel allocation profile, +.>Representing D2D offload execution costs;
setting a first constraint condition, including:
(1) The time delay of task execution cannot exceed the maximum time delay tolerance, and is expressed by the following formula:
(2) The task execution energy consumption cannot exceed the maximum energy consumption, which is desirably represented by the following formula:
(3) The user offload association constraint, expressed as a binary variable that establishes a D2D offload link D2D pairing between TDm and the SDs, must have an SDs service TDm occur as a D2D offload execution, expressed by the following equation:
δ ms ∈{0,1},
establishing a one-to-one user association constraint relation of D2D unloading, wherein the one-to-one user association constraint relation is expressed by the following formula:
(4) The communication resource allocation constraint is set such that one conventional cellular channel can be allocated at most one D2D offload link, and one D2D offload link can be multiplexed with at most one cellular channel, expressed by the following formula:
(5) Setting maximum power constraint, wherein the data transmission rate meets minimum rate constraint and data segmentation constraint conditions, and the method comprises the following specific steps:
wherein ,θm Representing TDm the amount of data the task is performing at D2D offload,representing the transmission power of links TDm to SDs in CDn-reusable channels, +.>Representing the signal-to-noise ratio of the D2D offload link.
Further, in step S2, a step is performed in which the method is constructed in a known θ m Andthe second optimization model P2 based on the offloading policy and the channel allocation is expressed as follows:
P2:
setting a second constraint condition, comprising:
further, constructing a vertex set of the hypergraph according to the set M, the set S and the set N, executing a second constraint condition, eliminating the hyperedge which does not meet the condition, reserving the hyperedge which meets the constraint condition to construct a feasible hypergraph, arranging the hyperedge in the feasible hypergraph in a descending order according to the weight of the hyperedge, and acquiring an initial independent set UA and an adjacent set UB through a greedy algorithm to construct a conflict graph, wherein the adjacent set UB is a set of all adjacent vertexes of the independent set UA;
searching claw in conflict graph according to search substitution algorithm, taking each node of UA in independent set in the conflict graph as central node, if a certain central node u i The overall task execution cost of the claw nodes in the corresponding adjacent set UB is smaller than that of the central node u i Then the claw node is used for replacing the central node u i And eliminating adjacent nodes between the independent set and the claw node to update the independent set UA, thereby completing the searching of the claw in the conflict graph and outputting the optimization result of the optimal independent set UA of the overall task execution cost, wherein the claw node is formed by a plurality of non-adjacent vertexes in the adjacent set UB, which conflict with each node of the independent set UA as a central node.
Further, at delta ms Andin the given case, the first network optimization model is converted into a third optimization model P3 based on power control and task allocation, wherein the third optimization model and constraints are as follows:
P3:
setting a third constraint condition, comprising:
in the third optimization model, deriving the boundary of the unloading fraction from the task execution time delay constraint condition and the task execution energy consumption constraint condition, namely obtaining theta by the delay threshold under the task execution time delay constraint condition m Is subject to energy consumption constraints to achieve θ m Upper boundary of (2); obtaining the time delay constraint condition of task execution, the energy consumption constraint condition of task execution and the constraint condition of data transmission rateAt this time, searching for the best +.> and />Thereby obtaining D2D computation offload result data.
Further, the iterative search algorithm is as follows:
input: τ m ，E max ，P max ，R min
And (3) outputting: θ m 、
Initializing parameters:
according to the constraint conditions: task execution time delay constraint conditions, task execution energy consumption constraint conditions and data segmentation constraint conditions are used for obtaining optimal by utilizing search algorithmIs->
wherein ,
iterative looping the above process until and />Outputting a result after convergence is finished, wherein tau m Representing maximum latency tolerance of task execution, I m Data size representing tasks in bits, < >>Representing the data rate of a D2D offload link, D m Representing the amount of computing resources required to complete a one-bit TDm task, f m Representing the computational capacity of each TDm, f s Representing the computational capacity of each SDs, +.>Representing the channel gain of TDm to SDs, < ->Representing the interference gain from cellular subscriber n to SDs, let +.>Representing the transmission power of cellular user N e N 0 Then it represents gaussian white noise, B represents channel bandwidth, η m Representing the energy consumption coefficient, η, of the task user device TDm s And the energy consumption coefficient of the idle user equipment SDs of the resource is represented.
Due to the adoption of the technical scheme, the application has the following beneficial effects: the D2D computation unloading method based on hypergraph matching computation and communication capacity enhancement disclosed by the application considers the coupling relation between communication and computation resources, performs joint optimization on task scheduling, unloading decision and resource allocation, and is used for reducing the system execution cost; the method comprises the following steps: firstly, establishing an optimized time delay and energy consumption weighted minimum cost optimization model, then decomposing an original problem into a plurality of sub-problems such as task proportion, power distribution, unloading decision and the like, describing the unloading decision and channel distribution as three-dimensional hypergraph matching, solving by utilizing a claw theorem, a conflict graph and other design heuristic algorithms, and solving by adopting quasi-projection for unloading proportion and power distribution; finally, simulation shows that the task execution cost of the mechanism is obviously reduced, and the spectrum efficiency and the calculation efficiency of the network are improved.
Drawings
FIG. 1 is a flow chart of a D2D computation offload method based on hypergraph matching computation and communication capacity enhancement of the present application;
FIG. 2 is a diagram of a D2D network system architecture of the present application;
FIG. 3 is a schematic diagram of a conflict graph architecture of the present application;
FIG. 4 is a diagram of a D2D computing offload simulation embodiment of the present application;
FIG. 5 is a diagram of the main set-up parameters in the simulation process of the present application;
FIG. 6 is a comparison of an embodiment of the present application based on task data size.
Detailed Description
The technical scheme of the present application and how the technical scheme of the present application solves the above technical problems are described in detail below with specific embodiments; the following embodiments may be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments; embodiments of the present application will now be described with reference to the accompanying drawings.
As shown in fig. 1, the embodiment of the application provides a D2D computation offload method based on hypergraph matching computation and communication capacity enhancement, which specifically includes the following steps:
acquiring channel state information and task state information to construct a first network optimization model, and setting constraint conditions of the first network optimization model;
as shown in fig. 2, the network system is composed of a base station and a plurality of user devices, and considering the calculation and communication full-load scenario, the base station is equipped with an edge server, but cannot provide more calculation unloading services, and the user can only process tasks through a local execution or D2D unloading mode; the user equipment in the network is divided into three types, namely task user equipment TD, service user equipment SD with idle resources and cellular user equipment CD, and if m= {1,2,., M } represents a TD set, then M is the number of TDs, and the mth TD is represented as TDm; SD is a set of resource-free user equipments with high computational power; let s= {1,2,..s } denote a set of SDs, where S is the number of SDs, and the S-th SD is the number of SDs; the CD is a cellular user that has communicated with the base station, and its channel is multiplexed for D2D offload data transmission; let n= {1,2,..n } denote the set of CDs, where N is the number of CDs and the nth CD is denoted CDn;
for a computing domain, a user selects local execution or D2D unloading execution according to factors such as task requirements, hardware conditions, channel states and the like; for the communication domain, two communication modes are included, one is traditional cellular communication, and the other is D2D offload communication, wherein the D2D offload communication carries out data transmission in an uplink spectrum multiplexing mode, therefore, a D2D receiving end receives transmission interference of a cellular user, a base station receives signal interference of a D2D transmitting end, and in order to avoid excessive same-frequency interference, one D2D pair can only multiplex one cellular channel, one cellular channel can only be multiplexed by one D2D pair, and the D2D links adopt orthogonal spectrum resources for communication;
in task execution mode, computational task usage of TDmTo describe, wherein I m Is the data size (in bits) of the task, D m Representing the amount of computing resources, τ, required to complete a one-bit TDm task m Is the task deadline, i.e., the maximum tolerable delay (seconds) for task execution;
the task execution modes include a local mode, a D2D offload mode, and a binary variable delta ms E {0,1}, M e M, S e S to represent the execution mode and association of the TD of the user; variable delta ms The task of =1 means TDm is offloaded to the SDs for processing over the D2D link, otherwise δ ms =0; when the tasks of the TD are in the fully local execution mode,when the task selection D2D offload of the TD is performed,/->In view of the limited resources of the terminal, one D2D SD may serve at most one offloading device, A TD can offload its calculation to a D2D SD at most, in which case the +.>
In the local execution mode, the computation delay is expressed asf m m.epsilon.M represents the computational capacity of each TDm, and the energy consumed by the corresponding local execution is expressed as +.> wherein ,ηm The energy consumption coefficient representing the task device TDm, which depends on the performance of the CPU, the values of the different devices being different; setting TDm execution cost as a weighted sum of task delay and energy consumption to evaluate task execution efficiency, expressed by the following formula:
U m ＝α t T m +α e E m
wherein ,αt ∈[0,1],α t ∈[0,1]Respectively represent weight coefficients corresponding to delay and energy consumption, and alpha t +α e =1, the weight coefficient can be set according to the user status and the requirement; when the user experiences a delay sensitive application, the user is more concerned about the delay and can set a e ＝0,α t When the user equipment is in a low battery state, the user prefers to save energy, and a can be set e ＝1,α t ＝0；
In the D2D execution mode, the D2D offload link multiplexes channel resources of the cellular user for communication with binary variablesTo represent the channel allocation profile; when->Indicating that a CDn is assigned to the offload link between TDm and SDs. The signal-to-noise ratio SINR of the D2D offload link is expressed as:
wherein ,channel gain, < > -representing TD m to SD s>Indicating the interference gain from cellular user n to SD s, let +.>Representing the transmission power of cellular user N e N +.>Representing transmission power from TD m to SD s, N 0 Then gaussian white noise is represented; and B represents the channel bandwidth, the data rate of the D2D offload link can be obtained according to Shannon's theorem as follows:
to ensure communication quality, the data rate of each D2D offload link should be greater thanIs a minimum threshold of (2);
based on the above, it can be known that the transmission delay of the task in the D2D offload mode isExecution delay ofThe total delay in D2D offload mode is:
correspondingly, in the D2D unloading mode, the energy consumption of task transmission isThe energy consumption of task execution isη s The energy consumption coefficient representing the idle user equipment SDs of the resource, the total energy consumption in the D2D unloading mode is:
in order to effectively utilize computing resources in a network, a partial offloading mode is adopted for task execution during D2D offloading, namely an application program is data partition oriented, wherein input bits of the task can be divided arbitrarily; introducing an offload fraction parameter θ m ∈[0，1]As a task partition variable, f s Representing the computational capacity of each SDs, more specifically θ m Task defined as TD m, 1- θ, data volume performed at D2D offload m Representing the local execution amount of the task of TD m;
since local computation can be performed simultaneously with the computation offload process, the overall delay of task execution is determined by the D2D delay or the longer local delay, and the overall delay of task execution in the D2D offload mode is:
the total energy consumption for task execution in D2D offload mode is:
correspondingly D2D offload execution cost is
Based on the above, a first network optimization model is constructed, and constraint conditions are set, wherein the first network optimization model P1 is represented as follows:
in the formula ,δms Representing a user association relationship between a task execution mode of the user TD and an unloading strategy; q (Q) m Representing the overall cost, including power consumption and time delay,representing channel allocation profile, +.>Representing D2D offload execution costs;
setting a first constraint condition, including:
(1) The time delay of task execution cannot exceed the maximum time delay tolerance, and is expressed by the following formula:
(2) The task execution energy consumption cannot exceed the maximum energy consumption, which is desirably represented by the following formula:
(3) The user uninstalls the association constraint, expressed as:
δ ms ∈{0,1},
the above expression is that establishing a D2D offload link D2D pairing between TDm and the SDs is a binary variable, and that an SDs service TDm must occur as a D2D offload execution;
the one-to-one user association constraints expressed as D2D offload;
(4) Communication resource allocation constraint conditions are set, one traditional cellular channel can be allocated with at most one D2D unloading link, and one D2D unloading link can be multiplexed with at most one cellular channel, as follows:
(5) Setting maximum power constraint, wherein the data transmission rate meets minimum rate constraint and data segmentation constraint conditions, and the method comprises the following specific steps:
wherein ,θm For representing the amount of data that TDm tasks are performed in D2D offload,representing the transmission power of links TDm to SDs in CDn-reusable channels, +.>Representing the signal-to-noise ratio of the D2D offload link;
in this embodiment, the first network optimization model is built through the above steps, and corresponding constraint conditions are set.
In the embodiment, model optimization is performed based on the network model, the data processing process is set to be a second optimization model based on an unloading strategy and channel allocation and a third optimization model based on power control and task allocation, and continuous optimization problems of power control and task segmentation are solved through a search method, so that maximization of system calculation and communication capacity is achieved; i.e. at a known angle θ m Andin the case of (a), converting the first network optimization model into a second optimization model P2 based on offloading policy and channel allocation, and at δ ms and />The first network optimization model is converted, in the given case, into a third optimization model P3 based on power control and task allocation.
Build at a known θ m Andthe second optimization model P2 based on the offloading policy and the channel allocation and the constraints are expressed as follows:
P2:
setting a second constraint condition, comprising:
in the embodiment, the second optimization model realizes how to perform task unloading by pairing the task node m and the idle node s, and how to select the spectrum resource of the cellular node n to perform task transmission, so that the overall task execution cost of the system is optimal, and model data optimization processing is performed by introducing three-dimensional hypergraph matching and utilizing a bipartite conflict graph theory and a heuristic algorithm;
based on the model, introducing a three-dimensional hypergraph, constructing a hypergraph vertex set by a task user set M, an idle user SDs set S and a cellular user CDs set N, and respectively forming a hyperedge E E by the combination of any vertex in the set M, S, N H = { M, S, N }, M e M, S e S, N e N, whose superside weight is represented by the task execution cost under the combination, i.e.Executing constraint conditions in the second optimization model P2, eliminating superedges which do not meet the constraint conditions, reserving the superedges which meet the constraint conditions to be feasible superedges, replacing each feasible superedge with a vertex to construct a feasible supergraph as shown in FIG. 3, arranging the superedges in the feasible supergraph in descending order according to the weights of the superedges, and obtaining an initial independent set U in the conflict graph through a greedy algorithm A And adjacent set U B ；
In the construction of conflict graphs, U is used A and UB To represent the vertex set U, U of the conflict graph A Representing independent sets in conflict graph, U B Is set U A The set of all adjacent vertices, U A ＝U/U B Vertex u i ∈U A And vertex u j ∈U B Having strong adjacency, u i And u is equal to j An edge exists between the two edges, and correspondingly, the superedge corresponding to the vertex has an intersecting relation;
the problem of minimum weight matching of the three-dimensional hypergraph is converted into the problem of finding the minimum weight independent set, in this scheme, the claws contained in the conflict graph need to be found out, in this embodimentWherein, each node in the independent set is taken as a central node and a plurality of non-adjacent vertexes which conflict with the central node, for a k-dimensional uniform hypergraph, at most k+1 non-intersecting neighbors are provided, in the three-dimensional uniform hypergraph, each central node is provided with at most three claws, the claws are searched in the conflict graph according to a searching substitution algorithm, if the whole task execution cost of the claw set corresponding to a certain central node is smaller than that of the central node, the central node is replaced by the claw node, and the adjacent nodes of the independent set and the claw node are removed, so that the UA of the independent set is updated, thereby completing the searching of the claws in the conflict graph and outputting the U of the best independent set of the whole task execution cost A So that the overall task execution cost of the conflict graph independent set will be optimal;
as shown in FIG. 3, independent set U A Node u in (a) 2 With two adjacent nodes u 5 and u7 Although u 4 There are five neighboring nodes u 5 ，u 6 ，u 7 ，u 8 ，u 9 But node u 5 and u6 Has the same elements, and therefore u 4 There are only three jaws at most. If a certain central node u i ∈U A Corresponding claw node set { G L_ i }∈U B If the overall task execution cost of the system is smaller than that of the central node, the claw node is used for replacing the central node u i And cull nodes adjacent to the claw node in the independent set, e.g. with node { u } 5 ,u 7 Substituted central node u 2 And will be combined with u 5 Adjacent node u 1 Delete, then original independent set { u } 1 ,u 2 ,u 3 ,u 4 Becomes { u } 5 ,u 7 ,u 3 ,u 4 The overall task execution cost of such an independent set of conflict graphs would be optimal.
At delta ms Andconverting the first network optimization model into a third optimization model P3 based on power control and task allocation, wherein the third optimization model and the constraint barsThe parts are as follows:
P3:
setting a third constraint condition, comprising:
in this embodiment, the third optimization model further performs result optimization on the output of the first optimization model based on the result of the second model, specifically how the implementation is specified at δ ms Andin the given case, search for the best offload fraction +.> and />The offload fraction plays an important role in task execution delay and energy to consume local execution and offload execution, and derives according to task execution delay constraint conditions and task execution energy consumption constraint conditionsOutputting the boundary of the unloading fraction, namely obtaining theta by the delay threshold under the constraint condition of task execution time delay m Is subject to energy consumption constraints to achieve θ m Is similarly obtained according to the task execution time delay constraint condition, the task execution energy consumption constraint condition and the data transmission rate constraint conditionAt this time, the optimal ++is found by using the iterative search algorithm> and />
The iterative search algorithm is as follows:
input: τ m ，E max ，P max ，R min
And (3) outputting: θ m 、
Initializing parameters:
according to the constraint conditions: task execution time delay constraint conditions, task execution energy consumption constraint conditions and data segmentation constraint conditions are used for obtaining optimal by utilizing search algorithmIs->
wherein ,
iterative looping the above process until and />Outputting a result after convergence is finished, wherein tau m Representing maximum latency tolerance of task execution, I m Data size representing tasks in bits, < >>Representing the data rate of a D2D offload link, D m Representing the amount of computing resources required to complete a one-bit TDm task, f m Representing the computational capacity of each TDm, f s Representing the computational capacity of each SDs, +.>Representing the channel gain of TDm to SDs, < ->Representing the interference gain from cellular subscriber n to SDs, let +.>Representing the transmission power of cellular user N e N 0 Then it represents gaussian white noise, B represents channel bandwidth, η m Representing the energy consumption coefficient, η, of the task user device TDm s And the energy consumption coefficient of the idle user equipment SDs of the resource is represented.
In this embodiment, a performance comparison of the joint computing offload and resource allocation algorithm proposed by the present application is given, in the local computing, each TDm computes its task locally, without offloading being involved; random unloading: each TDm randomly selects idle tasks to unload, and the proportion of task data to be unloaded is randomly determined; the communication and calculation resources are evenly distributed, and the selection of the idle nodes of the resources and the selection pair of the spectrum sharing nodes are randomly decided; JOAGT scheme: the scheme optimizes the proportion of task unloading data to minimize system delay without considering D2D pairing and resource allocation; mixing and unloading: each TD selects the best offload mode by comparing the local execution cost to the D2D offload cost, and the D2D offload user further selects a calculation and communication resource sharing strategy according to the three-dimensional matching algorithm presented herein.
For convenience, the joint computation offload and resource allocation based on the hypergraph algorithm proposed in this embodiment will be described as "JCRHG" in the figure.
In the simulation process, an underlying D2D unloading system consisting of one BS and a plurality of users is considered, wherein the system comprises task users, resource idle service users and cellular users; as shown in fig. 4, the size of the simulation area is 500m×500m, and the users are uniformly distributed in the simulation area; wherein, the shape represents task users, the circle represents service users with idle resources, the square represents cellular users, and the triangle represents base stations; the main setting parameters of the simulation are shown in a chart 5; our goal is to minimize the execution cost, which is a weighted sum of delay and energy consumption; therefore, we will analyze the three indexes in terms of task size, number of users, D2D communication distance, etc.; to better describe the various properties, the simulation results averaged over 1000 independent experiments.
The task execution cost of all algorithms increases with the increase of the size of the task input data, because the larger the size requirement of the task input data is, the higher the task execution delay is caused, the larger the energy consumption is, and the higher the task execution cost is caused; as shown in fig. 6, where the upper half of the histogram represents the delay and the lower half represents the energy consumption, it can be seen that the JCRHG offloading scheme has better performance; in particular, JCRHG costs about 5% less than JOAGT because JOAGT aims at minimizing delay, which may result in greater energy consumption; the costs of JCRHG are reduced by about 34% and 75% compared to the random offloading scheme and the local calculation scheme, respectively; both of these indicate that the proposed JCRHG scheme brings about the expected performance improvement.
From the above description of the embodiments, it will be apparent to those skilled in the art that the above embodiment method may be implemented by means of software plus necessary general hardware platform, or may be implemented by hardware, but in many cases the former is a preferred embodiment; based on such understanding, the technical solution of the present application may be embodied essentially or in part in the form of a computer software product stored on a storage medium (e.g. ROM/RAM, magnetic disk, optical disk) comprising instructions for causing a terminal (which may be a mobile phone, a computer, a server, or a network device, etc.) to perform the method of the embodiments of the present application.
The embodiments of the present application have been described above with reference to the accompanying drawings, but the present application is not limited to the above-described embodiments, which are merely illustrative and not restrictive, and many forms may be made by those having ordinary skill in the art without departing from the spirit of the present application and the scope of the claims, which are to be protected by the present application.

Claims (6)

1. A D2D computation unloading method based on hypergraph matching computation and communication capacity enhancement is characterized by comprising the following steps: comprises the following steps:
s1, obtaining channel state information and task state information to construct a first network optimization model, and setting a first constraint condition of the first network optimization model;
s2, performing optimal decoupling on the data processing result of the first network optimization model, generating a second optimization model based on an unloading strategy and channel allocation and a third optimization model based on power control and task allocation, and setting a second constraint condition and a third constraint condition based on the second optimization model and the third optimization model respectively;
s3, carrying out data analysis of unloading strategy and channel allocation in a second optimization model based on hypergraph matching, conflict graphs, independent set principles and second constraint conditions, and outputting an optimization result;
s4, in the third optimization model, based on the optimization result output by the second optimization model, performing data analysis by using an iterative search algorithm and a third constraint condition, and outputting D2D calculation unloading result data.
2. The hypergraph matching calculation and communication capacity enhancement-based D2D calculation offloading method of claim 1, wherein: in S1, a user equipment data set based on a first network optimization model is constructed, including a task user equipment TD, a service user equipment SD with idle resources, and a cellular user equipment CD, where m= {1,2,..m } represents a TD set, M is the number of TDs, and the mth TD is TDm; SD is a set of resource-free user equipments with high computational power; let s= {1,2,..s } denote a set of SDs, where S is the number of SDs, and S-th SD is denoted as SDs; the CD is a cellular user that has communicated with the base station, and the channel thereof is multiplexed by D2D offload data transmission, where n= {1,2,..n } represents a set of CDs, where N is the number of CDs, and the nth CD is denoted as CDn;
generating a first network optimization model, expressed by the following formula:
in the formula ,δms Representing a user association relationship between a task execution mode of the user TD and an unloading strategy; q (Q) m Representing user local executionThe overall cost of the row, including power consumption and time delay,representing channel allocation profile, +.>Representing D2D offload execution costs;
setting a first constraint condition, including:
(1) The time delay of task execution cannot exceed the maximum time delay tolerance, and is expressed by the following formula:
(2) The task execution energy consumption cannot exceed the maximum energy consumption, which is desirably represented by the following formula:
(3) The user offload association constraint, expressed as a binary variable that establishes a D2D offload link D2D pairing between TDm and the SDs, must have an SDs service TDm occur as a D2D offload execution, expressed by the following equation:
establishing a one-to-one user association constraint relation of D2D unloading, wherein the one-to-one user association constraint relation is expressed by the following formula:
(4) The communication resource allocation constraint is set such that one conventional cellular channel can be allocated at most one D2D offload link, and one D2D offload link can be multiplexed with at most one cellular channel, expressed by the following formula:
(5) Setting maximum power constraint, wherein the data transmission rate meets minimum rate constraint and data segmentation constraint conditions, and the method comprises the following specific steps:
wherein ,θm Representing TDm the amount of data the task is performing at D2D offload,representing the transmission power of links TDm to SDs in CDn-reusable channels, +.>Representing the signal-to-noise ratio of the D2D offload link.
3. The hypergraph matching calculation and communication capacity enhancement-based D2D calculation offloading method of claim 2, wherein: in step S2, the method is constructed in a known state theta m Andthe second optimization model P2 based on the offloading policy and the channel allocation is expressed as follows:
setting a second constraint condition, comprising:
4. the hypergraph matching calculation and communication capacity enhancement-based D2D calculation offload method of claim 3, characterized by: constructing a vertex set of the hypergraph according to the set M, the set S and the set N, executing a second constraint condition, eliminating the hyperedge which does not meet the constraint condition, reserving the hyperedge which meets the constraint condition to construct a feasible hypergraph, and according to the weight of the hyperedge, performing descending arrangement on the hyperedge in the feasible hypergraph, and acquiring an initial independent set UA and an adjacent set UB through a greedy algorithm to construct a conflict graph, wherein the adjacent set UB is a set of all adjacent vertexes of the independent set UA;
searching claw in conflict graph according to search substitution algorithm, taking each node of UA in independent set in the conflict graph as central node, if a certain central node u i The overall task execution cost of the claw nodes in the corresponding adjacent set UB is smaller than that of the central node u i Then the claw node is used for replacing the central node u i And eliminating adjacent nodes between the independent set and the claw node to update the independent set UA, thereby completing the searching of the claw in the conflict graph and outputting the optimization result of the optimal independent set UA of the overall task execution cost, wherein the claw node is formed by a plurality of non-adjacent vertexes in the adjacent set UB, which conflict with each node of the independent set UA as a central node.
5. The hypergraph matching calculation and communication capacity enhancement-based D2D calculation offloading method of claim 4, wherein: at delta ms Andin the given case, the first network optimization model is converted into a third optimization model P3 based on power control and task allocation, wherein the third optimization model and constraints are as follows:
setting a third constraint condition, comprising:
in the third optimization model, deriving the boundary of the unloading fraction from the task execution time delay constraint condition and the task execution energy consumption constraint condition, namely obtaining theta by the delay threshold under the task execution time delay constraint condition m Is subject to energy consumption constraints to achieve θ m Upper boundary of (2)The method comprises the steps of carrying out a first treatment on the surface of the Obtaining the time delay constraint condition of task execution, the energy consumption constraint condition of task execution and the constraint condition of data transmission rateAt this time, searching for the best +.> and />Thereby obtaining D2D computation offload result data.
6. The hypergraph matching calculation and communication capacity enhancement-based D2D calculation offloading method of claim 5, wherein: the iterative search algorithm is as follows:
input: τ m ，E max ，P max ，R min
And (3) outputting: θ m 、
Initializing parameters:
according to the constraint conditions: task execution time delay constraint conditions, task execution energy consumption constraint conditions and data segmentation constraint conditions are used for obtaining optimal by utilizing search algorithmIs->
wherein ,
iterative looping the above process until and />Outputting a result after convergence is finished, wherein tau m Representing maximum latency tolerance of task execution, I m Data size representing tasks in bits, < >>Representing the data rate of a D2D offload link, D m Representing the amount of computing resources required to complete a one-bit TDm task, f m Representing the computational capacity of each TDm, f s Representing the computational capacity of each SDs, +.>Representing the channel gain of TDm to SDs, < ->Representing the interference gain from cellular subscriber n to SDs, let +.>Representing the transmission power of cellular user N e N 0 Then it represents gaussian white noise, ">Representing channel bandwidth, eta m Representing the energy consumption coefficient, η, of the task user device TDm s And the energy consumption coefficient of the idle user equipment SDs of the resource is represented. />
CN202310628662.2A 2023-05-31 2023-05-31 D2D computing unloading method based on hypergraph matching computation and communication capacity enhancement Pending CN116634500A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310628662.2A CN116634500A (en) 2023-05-31 2023-05-31 D2D computing unloading method based on hypergraph matching computation and communication capacity enhancement

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310628662.2A CN116634500A (en) 2023-05-31 2023-05-31 D2D computing unloading method based on hypergraph matching computation and communication capacity enhancement

Publications (1)

Publication Number Publication Date
CN116634500A true CN116634500A (en) 2023-08-22

Family

ID=87641441

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310628662.2A Pending CN116634500A (en) 2023-05-31 2023-05-31 D2D computing unloading method based on hypergraph matching computation and communication capacity enhancement

Country Status (1)

Country Link
CN (1) CN116634500A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117440442A (en) * 2023-10-31 2024-01-23 重庆理工大学 Internet of things resource conflict-free distribution method and system based on graph reinforcement learning
CN117915481A (en) * 2024-01-18 2024-04-19 重庆理工大学 Resource allocation method and system of ultra-dense industrial Internet of things

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117440442A (en) * 2023-10-31 2024-01-23 重庆理工大学 Internet of things resource conflict-free distribution method and system based on graph reinforcement learning
CN117915481A (en) * 2024-01-18 2024-04-19 重庆理工大学 Resource allocation method and system of ultra-dense industrial Internet of things

Similar Documents

Publication Publication Date Title
Chen et al. Energy-efficient task offloading and resource allocation via deep reinforcement learning for augmented reality in mobile edge networks
Yang et al. DEBTS: Delay energy balanced task scheduling in homogeneous fog networks
CN109413724B (en) MEC-based task unloading and resource allocation scheme
CN111930436B (en) Random task queuing unloading optimization method based on edge calculation
CN108809695B (en) Distributed uplink unloading strategy facing mobile edge calculation
CN111314889A (en) Task unloading and resource allocation method based on mobile edge calculation in Internet of vehicles
CN111726826A (en) Online task unloading method in base station intensive edge computing network
CN116634500A (en) D2D computing unloading method based on hypergraph matching computation and communication capacity enhancement
Chen et al. Joint computation offloading and radio resource allocations in small-cell wireless cellular networks
CN112512065B (en) Method for unloading and migrating under mobile awareness in small cell network supporting MEC
Zhang et al. Mobility-aware cooperative task offloading and resource allocation in vehicular edge computing
Feng et al. Energy-efficient user selection and resource allocation in mobile edge computing
CN112491957A (en) Distributed computing unloading method and system under edge network environment
Ai et al. Dynamic offloading strategy for delay-sensitive task in mobile-edge computing networks
Zhu et al. Joint optimal allocation of wireless resource and MEC computation capability in vehicular network
Yaqub et al. Multi-objective resource optimization for hierarchical mobile edge computing
CN115866787A (en) Network resource allocation method integrating terminal direct transmission communication and multi-access edge calculation
Wang et al. Joint offloading decision and resource allocation in vehicular edge computing networks
Kaur et al. Intelligent spectrum management based on reinforcement learning schemes in cooperative cognitive radio networks
Cui et al. Multi-agent reinforcement learning for slicing resource allocation in vehicular networks
CN111343721B (en) D2D distributed resource allocation method for maximizing generalized energy efficiency of system
Wang et al. Offloading schemes in mobile edge computing with an assisted mechanism
Chen et al. Computation offloading and resource allocation based on cell-free radio access network
CN116193546A (en) Dynamic generalized user NOMA grouping CCHN-MEC network unloading decision optimization method
Zeng et al. Joint proportional task offloading and resource allocation for MEC in ultra-dense networks with improved whale optimization algorithm

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