CN112996073A - Wireless sensor low-power-consumption low-time-delay path type collaborative computing method - Google Patents

Wireless sensor low-power-consumption low-time-delay path type collaborative computing method Download PDF

Info

Publication number
CN112996073A
CN112996073A CN202110038750.8A CN202110038750A CN112996073A CN 112996073 A CN112996073 A CN 112996073A CN 202110038750 A CN202110038750 A CN 202110038750A CN 112996073 A CN112996073 A CN 112996073A
Authority
CN
China
Prior art keywords
energy consumption
node
mapping
task
omega
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.)
Granted
Application number
CN202110038750.8A
Other languages
Chinese (zh)
Other versions
CN112996073B (en
Inventor
任智源
王一鸣
程文驰
胡梅霞
陈晨
张海林
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.)
Xidian University
Original Assignee
Xidian University
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 Xidian University filed Critical Xidian University
Priority to CN202110038750.8A priority Critical patent/CN112996073B/en
Publication of CN112996073A publication Critical patent/CN112996073A/en
Application granted granted Critical
Publication of CN112996073B publication Critical patent/CN112996073B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W40/00—Communication routing or communication path finding
    • H04W40/02—Communication route or path selection, e.g. power-based or shortest path routing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00—Routing or path finding of packets in data switching networks
    • H04L45/12—Shortest path evaluation
    • H04L45/121—Shortest path evaluation by minimising delays
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00—Routing or path finding of packets in data switching networks
    • H04L45/14—Routing performance; Theoretical aspects
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/30—Services specially adapted for particular environments, situations or purposes
    • H04W4/38—Services specially adapted for particular environments, situations or purposes for collecting sensor information
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W40/00—Communication routing or communication path finding
    • H04W40/02—Communication route or path selection, e.g. power-based or shortest path routing
    • H04W40/04—Communication route or path selection, e.g. power-based or shortest path routing based on wireless node resources
    • H04W40/10—Communication route or path selection, e.g. power-based or shortest path routing based on wireless node resources based on available power or energy
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W84/00—Network topologies
    • H04W84/18—Self-organising networks, e.g. ad-hoc networks or sensor networks
    • 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)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The invention relates to a low-power-consumption low-time-delay path type collaborative computing method for a wireless sensor, which comprises the following steps: (1) constructing a WSN cloud network architecture; (2) formulating a task mapping strategy under the energy consumption constraint; (3) and solving the optimal mapping relation model by using a BPSO algorithm. The low-power-consumption low-time-delay path type collaborative computing method of the wireless sensor disclosed by the invention has the following beneficial effects: 1. based on a cloud and mist network architecture, a path type cooperative computing technology is introduced into a WSN, a mist computing layer is formed by using sink nodes in the WSN, a DAG-form delay sensitive service is mapped onto the sink nodes, and the computing power is used for carrying out step-by-step computation, so that the transmission and computation of the service are realized, and the service processing delay is reduced; 2. considering that the WSN node has limited energy consumption, the service must be completed under certain energy consumption constraint, so the energy consumption constraint is considered in the path type cooperative computing technology.

Description

Wireless sensor low-power-consumption low-time-delay path type collaborative computing method
Technical Field
The invention belongs to the field of communication, and particularly relates to a low-power-consumption low-time-delay path type collaborative computing method for a wireless sensor.
Background
With the continuous maturity of the technology of the internet of things and the large-scale construction of 5G wireless networks, the world of everything interconnection comes. As an important technical form of an internet of things underlying Network, a Wireless Sensor Network (WSN) is rapidly and widely developed, and has been successfully applied to multiple fields including military, natural environment monitoring, medical care, smart home, logistics tracking, and the like.
The existing WSN system generally comprises a sensor node, a sink node and a remote management center. The sensor nodes are generally integrated with one or more sensors of different types, and are mainly responsible for acquiring information of a monitoring area and transmitting acquired data to the sink nodes through one hop or multiple hops; the sink node is a backbone node of the WSN, has the main function of converging and forwarding network messages and has stronger data processing capacity and communication capacity; the remote management center is equivalent to a cloud platform, forms a resource pool by utilizing a network resource virtualization technology, stores and processes data sent by the WSN, and sends the processed data to a user. However, the communication overhead generated when the data acquired by the WSN is transmitted to the cloud computing center for computation is large, the time delay is high, and the time delay sensitive service such as military reconnaissance cannot be effectively supported.
Aiming at the problem of high transmission delay in a cloud computing mode, scholars propose an edge computing technology, and the propagation distance of data is shortened by segmenting the data and completing computation by utilizing the computing power of network edge equipment (such as a sink node in a WSN). However, the conventional edge calculation technology is not suitable for processing end-to-end services, and lacks the capability of calculation while transmitting.
Aiming at the problems of the traditional edge calculation, some researchers provide a path type collaborative calculation technology based on step segmentation. The technology adopts a Directed Acyclic Graph (DAG) model composed of a plurality of service functions to represent services, maps a task graph in a DAG form to a network graph, and utilizes the capability of network edge equipment to complete service processing in the data transmission process, thereby realizing transmission and calculation of the services and being more suitable for processing complex novel information services; and each network node only needs to load the service function mapped to the network node, so that the computing load of a single network device can be effectively reduced, and the method is suitable for WSN nodes with limited load capacity.
In view of the early research on the path-based cooperative computing technology, scholars represented traffic as a tree graph or a directed acyclic graph and mapped it onto a network in order to optimize the path computation rate of the network. The ist and other scholars provide a scheme for mapping the edges and the nodes of the task graph to the network topology graph simultaneously while considering the node constraint capability, and solve the optimal task scheduling scheme in the distributed cloud network scene. Michael et al formulated a task graph mapping scheme with branches to extend the task graph to an arbitrary DAG model.
However, the above researches on path-based collaborative computing do not consider the problem of energy consumption of nodes, the increase of tasks processed by the WSN leads to the increase of overall energy consumption of the network, and shortens the service life of the WSN nodes, and the WSN nodes are usually located in the field, are limited in energy and difficult to supplement, so how to perform task mapping by path-based collaborative computing is how to minimize service processing delay under the premise of energy consumption constraint, and becomes a technical problem which needs to be solved urgently.
Disclosure of Invention
The purpose of the invention is as follows: aiming at the problem of high transmission delay when cloud computing is applied to a wireless sensor network delay sensitive service, a low-power-consumption low-delay path type collaborative computing method of a wireless sensor is provided. The method is based on a cloud and fog network architecture, and the cloud and fog network architecture utilizes a sink node to form a fog computing layer; task calculation is completed in steps based on the calculation capability of the fog calculation layer in the data transmission process, and the task processing time delay is reduced; due to the fact that the computing power of the sink node is weak, energy consumption is increased due to time delay reduction, the service life of a wireless sensor system is shortened, a task mapping strategy under energy consumption constraint is provided for the purpose, and a time delay Optimization problem under energy consumption constraint is solved by using a Binary Particle Swarm Optimization (BPSO) algorithm.
The technical scheme is as follows: the low-power-consumption low-time-delay path type collaborative computing method for the wireless sensor comprises the following steps:
(1) constructing a WSN cloud network architecture;
(2) and formulating a task mapping strategy under the constraint of energy consumption:
mapping a directed acyclic graph G in a DAG form to a fog network of an undirected connected graph U based on the WSN cloud and fog network architecture obtained in the step (1), and constructing an optimal mapping relation model from the directed acyclic graph G to the undirected connected graph U;
(3) and (3) solving the optimal mapping relation model obtained in the step (2) by using a BPSO algorithm.
Further, the WSN cloud and fog network architecture in step (1) includes a sensing layer, a fog computing layer, and a cloud computing layer from bottom to top, wherein:
the sensing layer consists of wireless sensors integrated with one or more types of sensors and is used for monitoring the deployment area;
the fog computing layer is composed of a plurality of sink nodes with data processing capacity and communication capacity, the sink nodes are communicated and interconnected with the wireless sensor, and the fog computing layer is used for forwarding and processing data generated by the sensing layer;
the cloud computing layer is composed of a plurality of server clusters, and the server clusters are connected with the sink nodes in a communication interconnection mode through communication links and used for monitoring and managing the WSN cloud network architecture.
Further, the step (2) comprises the following steps:
(21) constructing a mapping rule model from a directed acyclic graph G to an undirected connected graph U in a DAG form:
in a mapping rule model from a directed acyclic graph G in DAG format to an undirected connected graph U, the directed acyclic graph G ═ (Ω, Γ) represents a task model, and Ω ═ { ω ═ ω is defined1,ω2,…,ωs,ωs+1,…,ωl-1,ωl| s is more than or equal to 1, l is more than s +1} is a node set of G, wherein:
ω1,ω2,…,ωsis s task starting points, ωs+1,…,ωl-1Is an intermediate task node, omegalIs the task end point;
Γ is the set of directed edges of G, defining Φ↑(ωi)＝{ωj|(ωj,ωi) E is gamma is omegaiThe forward node set of (2);
in addition, the WSN topology is represented by an undirected connectivity graph U ═ (V, K), and V ═ ν is defined1,ν2,…,νs,νs+1,…,νt-1,νt| s is more than or equal to 1, t is more than s +1} is a node set of U, wherein v1,ν2,…,νsFinger service initiating node, vs+1,…,νt-1Being a relay node, vtThe nodes are directly connected with the users;
k is a set of U edges, each edge supporting bidirectional data transmission, using
Figure RE-GDA0003062872970000041
To represent node viTo vjThe shortest path of (2);
definition of
Figure RE-GDA0003062872970000051
Is a shortest path set;
Figure RE-GDA0003062872970000052
the data forwarding nodes passing through the shortest path are collected;
Figure RE-GDA0003062872970000053
to the slave node viTo vjTime delay for transmitting unit data amount along shortest path;
network edge transmission rate of graph U
Figure RE-GDA0003062872970000054
The connection relation of the sum node is used as input and can be obtained through a Floyed algorithm
Figure RE-GDA0003062872970000055
The mapping rule from the directed acyclic graph G to the directed connected graph U is as follows:
define 1. the mapping rule of Ω to V is ε: Ω → V, and ε should satisfy the condition of formula (1):
Figure RE-GDA0003062872970000056
epsilon will be omega task starting point omega1,ω2,…,ωsV-mapped task initiating node V1,ν2,…,νs(ii) a Intermediate task node omegas+1,…,ωl-1Mapping to arbitrary relay node vs+1,…,νt-1(ii) a Will task end point omegalMapping as a node v directly connected to a usert；
Define 2. the mapping of Γ to P is Γ → P, and γ needs to satisfy the condition of formula (2):
Figure RE-GDA0003062872970000057
y maps the directed edges in set Γ to node epsilon (ω) in graph Ui) To epsilon (omega)j) Shortest path of
Figure RE-GDA0003062872970000058
(22) Constructing a time delay model based on the mapping rule model obtained in the step (21):
subtask omegaiAt a certain timeThe time delay in the mapping relationship can be expressed as equation (3):
Figure RE-GDA0003062872970000059
wherein:
Figure RE-GDA00030628729700000510
to proceed to the subtask omegaiCumulative time delay of time;
Figure RE-GDA00030628729700000511
is omegaiCalculating time delay;
Figure RE-GDA0003062872970000061
is a node ε (ω)i) The computing power of (a);
alpha is a task computation complexity coefficient;
the task processing delay of the directed acyclic graph G is the task end point omegalThe delay of (2) is as shown in equation (4):
T(G)＝T(ωl) (4)
(23) constructing an energy consumption model based on the mapping rule model obtained in the step (21) and the time delay model obtained in the step (22), wherein:
network node viIs equal to network node viThe sum of idle energy consumption and active energy consumption;
(231) idle energy consumption
(2311) Mapping node ε (ω)i) The idle energy consumption is as shown in equation (5):
Figure RE-GDA0003062872970000062
wherein the content of the first and second substances,
Figure RE-GDA0003062872970000063
meaning epsilon (omega)i) Power in idle state;
Figure RE-GDA0003062872970000064
meaning epsilon (omega)i) Time in idle state in a certain task;
Figure RE-GDA0003062872970000065
the calculation time, the data sending time and the data receiving time are respectively, and the three times do not coincide, and the calculation formula is shown in formulas (6) to (8):
Figure RE-GDA0003062872970000066
Figure RE-GDA0003062872970000067
Figure RE-GDA0003062872970000068
from formulas (5) to (8) to obtain ε (ω)i) The idle energy consumption is as follows (9):
Figure RE-GDA0003062872970000071
(2312) forwarding node
Figure RE-GDA0003062872970000072
The idle energy consumption is as shown in equation (10):
Figure RE-GDA0003062872970000073
Figure RE-GDA0003062872970000074
the calculation can be performed using equations (11) to (12):
Figure RE-GDA0003062872970000075
Figure RE-GDA0003062872970000076
from (10) to (12)
Figure RE-GDA0003062872970000077
The idle energy consumption is as shown in equation (13):
Figure RE-GDA0003062872970000078
(232) and active energy consumption:
the activity energy consumption comprises the following calculation energy consumption and transmission energy consumption:
(2321) calculating energy consumption:
the calculation energy consumption is only determined by the mapping node epsilon (omega)i) Yield, as in formula (14):
Figure RE-GDA0003062872970000081
where k > 0 and σ ≧ 2 are both positive real numbers, σ and k are set to 3 and 10, respectively-28；
(2322) And energy consumption in transmission:
mapping node ε (ω)i) The transmission energy of (A) is as follows:
Figure RE-GDA0003062872970000082
Figure RE-GDA0003062872970000083
wherein, PTAnd PRThe transmit power and the receive power of the node, respectively, so that e (ω)i) The activity energy consumption of (2) is as follows:
Figure RE-GDA0003062872970000084
forwarding node
Figure RE-GDA0003062872970000085
Includes only the transmission energy consumption, as in equation (18):
Figure RE-GDA0003062872970000086
from the equations (9) and (17), the mapping node ε (ω)i) The total energy consumption of (2) is as follows:
Figure 100002_3
obtained by the equations (13) and (18), forwarding node
Figure RE-GDA0003062872970000092
The total energy consumption of (a) is as follows:
Figure 2
the total energy consumption of the entire mist network is shown as equation (21):
Figure RE-GDA0003062872970000094
let the maximum energy contained in the whole fog network be EmaxThen, in a certain task G, the energy consumption generated by the network needs to be less than or equal to the maximum energy consumption, as shown in equation (22):
Figure RE-GDA0003062872970000101
(24) constructing an optimization model of the mapping rule from the directed acyclic graph G in the DAG form to the undirected connected graph U, and based on the steps (21) and (23), providing the mapping rule from the directed acyclic graph G in the DAG form to the undirected connected graph U, optimizing the model, and establishing a binary optimization problem:
define 3. subtask node ωpAnd fog network node vqThe mapping relationship of (1) is as follows: when in use
Figure RE-GDA0003062872970000102
When it is, i.e. ωpIs mapped as vq(ii) a When in use
Figure RE-GDA0003062872970000103
Time, omegapWill not be mapped as vqThen, then
Figure RE-GDA0003062872970000104
Satisfies formula (23):
Figure RE-GDA0003062872970000105
based on the definition 3, ωpTo vqCan be constructed as a mapping matrix X of l X t, as in equation (24):
Figure RE-GDA0003062872970000106
the subtask ω represented by equation (5)iCan be expressed by equation (25):
Figure RE-GDA0003062872970000107
the latency of task G may be expressed as a function of X, as in equation (26):
T(G)＝F(X) (26)
formula (19) isExpressed mapping node ε (ω)i) Can be expressed by the formula (27):
Figure RE-GDA0003062872970000111
the forwarding node represented by the formula (20)
Figure RE-GDA0003062872970000112
Can be expressed by equation (28):
Figure RE-GDA0003062872970000113
then, the optimal mapping relationship from the directed acyclic graph G to the undirected connected graph U under the energy consumption constraint is modeled as follows:
X＝arg min(F(X))
Figure RE-GDA0003062872970000121
further, the BPSO algorithm in step (3) is mainly used for optimizing the constraint problem of the discrete space, and limits the position of the particle to 0 or 1, and is applicable to the binary optimization problem proposed by equation (29):
when BPSO algorithm is adopted, particle swarm
Figure RE-GDA0003062872970000122
Moving within the search space I to find the best position,
Figure RE-GDA0003062872970000123
in NmaxFor maximum iteration number, M is the particle swarm size, N is equal to {1,2, …, NmaxThe iteration times are;
in the nth iteration, the position and velocity of the ith particle may be expressed as:
Figure RE-GDA0003062872970000124
Figure RE-GDA0003062872970000125
in the formula (30), Xn(i)∈I，
Figure RE-GDA0003062872970000126
In the formula (31), Vn(i)∈O，
Figure RE-GDA0003062872970000127
For the ith particle, in the nth iteration, the velocity update formula is as follows (32):
Figure RE-GDA0003062872970000128
in the formula (32), the compound represented by the formula (32),
Figure RE-GDA0003062872970000129
and
Figure RE-GDA00030628729700001210
respectively the local and global optimum positions of the particle, w is the inertial weight, gamma1And gamma2As an acceleration factor, beta1And beta2Is in the interval of [0,1 ]]Random numbers uniformly distributed therein;
the position update formula of the BPSO algorithm is as shown in formulas (33) to (34):
Figure RE-GDA0003062872970000131
Figure RE-GDA0003062872970000132
the fitness function of the algorithm is as follows (35):
f(X)＝T(G)＝F(X) (35)。
has the advantages that: the low-power-consumption low-time-delay path type collaborative computing method of the wireless sensor disclosed by the invention has the following beneficial effects:
1. based on a cloud and mist network architecture, a path type cooperative computing technology is introduced into a WSN, a mist computing layer is formed by using sink nodes in the WSN, a DAG-form delay sensitive service is mapped onto the sink nodes, and the computing power is used for carrying out step-by-step computation, so that the transmission and computation of the service are realized, and the service processing delay is reduced;
2. considering that the WSN node has limited energy consumption, the service must be completed under certain energy consumption constraint, so the energy consumption constraint is considered in the path type cooperative computing technology.
Drawings
Fig. 1 is a flowchart of a low-power-consumption low-delay path-type cooperative computing method for a wireless sensor disclosed by the invention.
Fig. 2 is a schematic diagram of a WSN cloud network architecture.
FIG. 3 is a schematic diagram of a mapping of a directed acyclic graph G to an undirected connected graph U, wherein:
1-cloud computing layer
11-communication link
2-fog calculating layer
21-sink node
22-user
3-sensing layer
31-Wireless sensor
32-deployment area
The specific implementation mode is as follows:
the following describes in detail specific embodiments of the present invention.
As shown in fig. 1, the low-power-consumption low-delay path type collaborative calculation method for the wireless sensor includes the following steps:
(1) constructing a WSN cloud network architecture;
(2) and formulating a task mapping strategy under the constraint of energy consumption:
mapping a directed acyclic graph G in a DAG form to a fog network of an undirected connected graph U based on the WSN cloud and fog network architecture obtained in the step (1), and constructing an optimal mapping relation model from the directed acyclic graph G to the undirected connected graph U;
(3) and (3) solving the optimal mapping relation model obtained in the step (2) by using a BPSO algorithm.
Further, as shown in fig. 2, the WSN cloud and fog network architecture in step (1) includes, from bottom to top, a sensing layer 3, a fog computing layer 2, and a cloud computing layer 1, where:
the sensing layer 3 is composed of a wireless sensor 31 integrated with one or more types of sensors and used for monitoring a deployment area 32;
the fog computing layer 2 is composed of a plurality of sink nodes 21 with data processing capacity and communication capacity, the sink nodes 21 are in communication interconnection with the wireless sensor, and the fog computing layer 2 is used for forwarding and processing data generated by a user 22 in the sensing layer 3;
the cloud computing layer 1 is composed of a plurality of server clusters, and the server clusters are connected with the sink node 21 in a communication interconnection mode through the communication link 11 and used for monitoring and managing the WSN cloud network architecture.
Further, the step (2) comprises the following steps:
(21) constructing a mapping rule model from a directed acyclic graph G to an undirected connected graph U in a DAG form:
in a mapping rule model from a directed acyclic graph G in DAG format to an undirected connected graph U, the directed acyclic graph G ═ (Ω, Γ) represents a task model, and Ω ═ { ω ═ ω is defined1,ω2,…,ωs,ωs+1,…,ωl-1,ωl| s is more than or equal to 1, l is more than s +1} is a node set of G, wherein:
ω1,ω2,…,ωsis s task starting points, ωs+1,…,ωl-1Is an intermediate task node, omegalIs the task end point;
Γ is the set of directed edges of G, defining Φ↑(ωi)＝{ωj|(ωj,ωi) E is gamma is omegaiThe forward node set of (2);
in addition, the WSN topology is represented by an undirected connectivity graph U ═ (V, K), and V ═ ν is defined1,ν2,…,νs,νs+1,…,νt-1,νt| s is more than or equal to 1, t is more than s +1} is a node set of U, wherein v1,ν2,…,νsFinger service initiating node, vs+1,…,νt-1Being a relay node, vtThe nodes are directly connected with the users;
k is a set of U edges, each edge supporting bidirectional data transmission, using
Figure RE-GDA0003062872970000151
To represent node viTo vjThe shortest path of (2);
definition of
Figure RE-GDA0003062872970000152
Is a shortest path set;
Figure RE-GDA0003062872970000153
the data forwarding nodes passing through the shortest path are collected;
Figure RE-GDA0003062872970000161
to the slave node viTo vjTime delay for transmitting unit data amount along shortest path;
network edge transmission rate of graph U
Figure RE-GDA0003062872970000162
The connection relation of the sum node is used as input and can be obtained through a Floyed algorithm
Figure RE-GDA0003062872970000163
The mapping rule from the directed acyclic graph G to the directed connected graph U is as follows:
define 1. the mapping rule of Ω to V is ε: Ω → V, and ε should satisfy the condition of formula (1):
Figure RE-GDA0003062872970000164
epsilon will be omega task starting point omega1,ω2,…,ωsV-mapped task initiating node V1,ν2,…,νs(ii) a Intermediate task node omegas+1,…,ωl-1Mapping to arbitrary relay node vs+1,…,νt-1(ii) a Will task end point omegalMapping as a node v directly connected to a usert；
Define 2. the mapping of Γ to P is Γ → P, and γ needs to satisfy the condition of formula (2):
Figure RE-GDA0003062872970000165
y maps the directed edges in set Γ to node epsilon (ω) in graph Ui) To epsilon (omega)j) Shortest path of
Figure RE-GDA0003062872970000166
(22) Constructing a time delay model based on the mapping rule model obtained in the step (21):
subtask omegaiThe time delay in a certain mapping can be expressed as equation (3):
Figure RE-GDA0003062872970000167
wherein:
Figure RE-GDA0003062872970000168
to proceed to the subtask omegaiCumulative time delay of time;
Figure RE-GDA0003062872970000169
is omegaiCalculating time delay;
Figure RE-GDA00030628729700001610
is a node ε (ω)i) The computing power of (a);
alpha is a task computation complexity coefficient;
the task processing delay of the directed acyclic graph G is the task end point omegalThe delay of (2) is as shown in equation (4):
T(G)＝T(ωl) (4)
(23) constructing an energy consumption model based on the mapping rule model obtained in the step (21) and the time delay model obtained in the step (22), wherein:
network node viIs equal to network node viThe sum of idle energy consumption and active energy consumption;
(231) idle energy consumption
(2311) Mapping node ε (ω)i) The idle energy consumption is as shown in equation (5):
Figure RE-GDA0003062872970000171
wherein the content of the first and second substances,
Figure RE-GDA0003062872970000172
meaning epsilon (omega)i) Power in idle state;
Figure RE-GDA0003062872970000173
meaning epsilon (omega)i) Time in idle state in a certain task;
Figure RE-GDA0003062872970000174
the calculation time, the data sending time and the data receiving time are respectively, and the three times do not coincide, and the calculation formula is shown in formulas (6) to (8):
Figure RE-GDA0003062872970000175
Figure RE-GDA0003062872970000176
Figure RE-GDA0003062872970000177
from formulas (5) to (8) to obtain ε (ω)i) The idle energy consumption is as follows (9):
Figure RE-GDA0003062872970000181
(2312) forwarding node
Figure RE-GDA0003062872970000182
The idle energy consumption is as shown in equation (10):
Figure RE-GDA0003062872970000183
Figure RE-GDA0003062872970000184
the calculation can be performed using equations (11) to (12):
Figure RE-GDA0003062872970000185
Figure RE-GDA0003062872970000186
from (10) to (12)
Figure RE-GDA0003062872970000187
The idle energy consumption is as shown in equation (13):
Figure RE-GDA0003062872970000188
(232) and active energy consumption:
the activity energy consumption comprises the following calculation energy consumption and transmission energy consumption:
(2321) calculating energy consumption:
the calculation energy consumption is only determined by the mapping node epsilon (omega)i) Yield, as in formula (14):
Figure RE-GDA0003062872970000191
where k > 0 and σ ≧ 2 are both positive real numbers, σ and k are set to 3 and 10, respectively-28；
(2322) And energy consumption in transmission:
mapping node ε (ω)i) The transmission energy of (A) is as follows:
Figure RE-GDA0003062872970000192
Figure RE-GDA0003062872970000193
wherein, PTAnd PRThe transmit power and the receive power of the node, respectively, so that e (ω)i) The activity energy consumption of (2) is as follows:
Figure RE-GDA0003062872970000194
forwarding node
Figure RE-GDA0003062872970000195
Includes only the transmission energy consumption, as in equation (18):
Figure RE-GDA0003062872970000196
mapping the nodes obtained by the equations (9) and (17)ε(ωi) The total energy consumption of (2) is as follows:
Figure 4
obtained by the equations (13) and (18), forwarding node
Figure RE-GDA0003062872970000202
The total energy consumption of (a) is as follows:
Figure 5
the total energy consumption of the entire mist network is shown as equation (21):
Figure RE-GDA0003062872970000204
let the maximum energy contained in the whole fog network be EmaxThen, in a certain task G, the energy consumption generated by the network needs to be less than or equal to the maximum energy consumption, as shown in equation (22):
Figure RE-GDA0003062872970000211
(24) constructing an optimization model of the mapping rule from the directed acyclic graph G in the DAG form to the undirected connected graph U, and based on the steps (21) and (23), providing the mapping rule from the directed acyclic graph G in the DAG form to the undirected connected graph U, optimizing the model, and establishing a binary optimization problem:
define 3. subtask node ωpAnd fog network node vqThe mapping relationship of (1) is as follows: when in use
Figure RE-GDA0003062872970000212
When it is, i.e. ωpIs mapped as vq(ii) a When in use
Figure RE-GDA0003062872970000213
Time, omegapWill not be mapped as vqThen, then
Figure RE-GDA0003062872970000214
Satisfies formula (23):
Figure RE-GDA0003062872970000215
based on the definition 3, ωpTo vqCan be constructed as a mapping matrix X of l X t, as in equation (24):
Figure RE-GDA0003062872970000216
the subtask ω represented by equation (5)iCan be expressed by equation (25):
Figure RE-GDA0003062872970000217
the latency of task G may be expressed as a function of X, as in equation (26):
T(G)＝F(X) (26)
the mapping node ε (ω) represented by formula (19)i) Can be expressed by the formula (27):
Figure RE-GDA0003062872970000221
the forwarding node represented by the formula (20)
Figure RE-GDA0003062872970000222
Can be expressed by equation (28):
Figure RE-GDA0003062872970000223
then, the optimal mapping relationship from the directed acyclic graph G to the undirected connected graph U under the energy consumption constraint is modeled as follows:
Figure RE-GDA0003062872970000231
further, the BPSO algorithm in step (3) is mainly used for optimizing the constraint problem of the discrete space, and limits the position of the particle to 0 or 1, and is applicable to the binary optimization problem proposed by equation (29):
when BPSO algorithm is adopted, particle swarm
Figure RE-GDA0003062872970000232
Moving within the search space I to find the best position,
Figure RE-GDA0003062872970000233
in NmaxFor maximum iteration number, M is the particle swarm size, N is equal to {1,2, …, NmaxThe iteration times are;
in the nth iteration, the position and velocity of the ith particle may be expressed as:
Figure RE-GDA0003062872970000234
Figure RE-GDA0003062872970000235
in the formula (30), Xn(i)∈I，
Figure RE-GDA0003062872970000236
In the formula (31), Vn(i)∈O，
Figure RE-GDA0003062872970000237
For the ith particle, in the nth iteration, the velocity update formula is as follows (32):
Figure RE-GDA0003062872970000238
in the formula (32), the compound represented by the formula (32),
Figure RE-GDA0003062872970000239
and
Figure RE-GDA00030628729700002310
respectively the local and global optimum positions of the particle, w is the inertial weight, gamma1And gamma2As an acceleration factor, beta1And beta2Is in the interval of [0,1 ]]Random numbers uniformly distributed therein;
the position update formula of the BPSO algorithm is as shown in formulas (33) to (34):
Figure RE-GDA0003062872970000241
Figure RE-GDA0003062872970000242
the fitness function of the algorithm is as follows (35):
f(X)＝T(G)＝F(X) (35)。
the specific BPSO algorithm is shown below:
Figure RE-GDA0003062872970000243
the embodiments of the present invention have been described in detail. However, the present invention is not limited to the above-described embodiments, and various changes can be made within the knowledge of those skilled in the art without departing from the spirit of the present invention.

Claims (4)

1. The low-power-consumption low-time-delay path type collaborative computing method for the wireless sensor is characterized by comprising the following steps of:
(1) constructing a WSN cloud network architecture;
(2) and formulating a task mapping strategy under the constraint of energy consumption:
mapping a directed acyclic graph G in a DAG form to a fog network of an undirected connected graph U based on the WSN cloud and fog network architecture obtained in the step (1), and constructing an optimal mapping relation model from the directed acyclic graph G to the undirected connected graph U;
(3) and (3) solving the optimal mapping relation model obtained in the step (2) by using a BPSO algorithm.
2. The wireless sensor low-power-consumption low-latency path-based collaborative computing method of claim 1, wherein the WSN cloud network architecture in step (1) comprises a sensing layer, a fog computing layer and a cloud computing layer from bottom to top, wherein:
the sensing layer consists of wireless sensors integrated with one or more types of sensors and is used for monitoring the deployment area;
the fog computing layer is composed of a plurality of sink nodes with data processing capacity and communication capacity, the sink nodes are communicated and interconnected with the wireless sensor, and the fog computing layer is used for forwarding and processing data generated by the sensing layer;
the cloud computing layer is composed of a plurality of server clusters, and the server clusters are connected with the sink nodes in a communication interconnection mode through communication links and used for monitoring and managing the WSN cloud network architecture.
3. The wireless sensor low-power-consumption low-time-delay path type collaborative computing method according to claim 1, wherein the step (2) comprises the following steps:
(21) constructing a mapping rule model from a directed acyclic graph G to an undirected connected graph U in a DAG form:
in a mapping rule model from a directed acyclic graph G in DAG format to an undirected connected graph U, the directed acyclic graph G ═ (Ω, Γ) represents a task model, and Ω ═ { ω ═ ω is defined1,ω2,…,ωs,ωs+1,…,ωl-1,ωl| s is more than or equal to 1, l is more than s +1} is a node set of G, wherein:
ω1,ω2,…,ωsis s task starting points, ωs+1,…,ωl-1Is an intermediate taskNode, ωlIs the task end point;
Γ is the set of directed edges of G, defining Φ↑(ωi)＝{ωj|(ωj,ωi) E is gamma is omegaiThe forward node set of (2);
in addition, the WSN topology is represented by an undirected connectivity graph U ═ (V, K), and V ═ ν is defined1,ν2,…,νs,νs+1,…,νt-1,νt| s is more than or equal to 1, t is more than s +1} is a node set of U, wherein v1,ν2,…,νsFinger service initiating node, vs+1,…,νt-1Being a relay node, vtThe nodes are directly connected with the users;
k is a set of U edges, each edge supporting bidirectional data transmission, using
Figure RE-FDA0003062872960000021
To represent node viTo vjThe shortest path of (2);
definition of
Figure RE-FDA0003062872960000022
Is a shortest path set;
Figure RE-FDA0003062872960000023
the data forwarding nodes passing through the shortest path are collected;
Figure RE-FDA0003062872960000024
to the slave node viTo vjTime delay for transmitting unit data amount along shortest path;
network edge transmission rate of graph U
Figure RE-FDA0003062872960000025
The connection relation of the sum node is used as input and can be obtained through a Floyed algorithm
Figure RE-FDA0003062872960000026
The mapping rule from the directed acyclic graph G to the directed connected graph U is as follows:
define 1. the mapping rule of Ω to V is ε: Ω → V, and ε should satisfy the condition of formula (1):
Figure RE-FDA0003062872960000031
epsilon will be omega task starting point omega1,ω2,…,ωsV-mapped task initiating node V1,ν2,…,νs(ii) a Intermediate task node omegas+1,…,ωl-1Mapping to arbitrary relay node vs+1,…,νt-1(ii) a Will task end point omegalMapping as a node v directly connected to a usert；
Define 2. the mapping of Γ to P is Γ → P, and γ needs to satisfy the condition of formula (2):
Figure RE-FDA0003062872960000032
y maps the directed edges in set Γ to node epsilon (ω) in graph Ui) To epsilon (omega)j) Shortest path of
Figure RE-FDA0003062872960000033
(22) Constructing a time delay model based on the mapping rule model obtained in the step (21):
subtask omegaiThe time delay in a certain mapping can be expressed as equation (3):
Figure RE-FDA0003062872960000034
wherein:
Figure RE-FDA0003062872960000035
to proceed to the subtask omegaiCumulative time delay of time;
Figure RE-FDA0003062872960000036
is omegaiCalculating time delay;
Figure RE-FDA0003062872960000037
is a node ε (ω)i) The computing power of (a);
alpha is a task computation complexity coefficient;
the task processing delay of the directed acyclic graph G is the task end point omegalThe delay of (2) is as shown in equation (4):
T(G)＝T(ωl) (4)
(23) constructing an energy consumption model based on the mapping rule model obtained in the step (21) and the time delay model obtained in the step (22), wherein:
network node viIs equal to network node viThe sum of idle energy consumption and active energy consumption;
(231) idle energy consumption
(2311) Mapping node ε (ω)i) The idle energy consumption is as shown in equation (5):
Figure RE-FDA0003062872960000041
wherein the content of the first and second substances,
Figure RE-FDA0003062872960000042
meaning epsilon (omega)i) Power in idle state;
Figure RE-FDA0003062872960000043
meaning epsilon (omega)i) Time in idle state in a certain task;
Figure RE-FDA0003062872960000044
the calculation time, the data sending time and the data receiving time are respectively, and the three times do not coincide, and the calculation formula is shown in formulas (6) to (8):
Figure RE-FDA0003062872960000045
Figure RE-FDA0003062872960000046
Figure RE-FDA0003062872960000047
from formulas (5) to (8) to obtain ε (ω)i) The idle energy consumption is as follows (9):
Figure RE-FDA0003062872960000048
(2312) forwarding node
Figure RE-FDA0003062872960000049
The idle energy consumption is as shown in equation (10):
Figure RE-FDA0003062872960000051
Figure RE-FDA0003062872960000052
the calculation can be performed using equations (11) to (12):
Figure RE-FDA0003062872960000053
Figure RE-FDA0003062872960000054
from (10) to (12)
Figure RE-FDA0003062872960000055
The idle energy consumption is as shown in equation (13):
Figure RE-FDA0003062872960000056
(232) and active energy consumption:
the activity energy consumption comprises the following calculation energy consumption and transmission energy consumption:
(2321) calculating energy consumption:
the calculation energy consumption is only determined by the mapping node epsilon (omega)i) Yield, as in formula (14):
Figure RE-FDA0003062872960000057
where k > 0 and σ ≧ 2 are both positive real numbers, σ and k are set to 3 and 10, respectively-28；
(2322) And energy consumption in transmission:
mapping node ε (ω)i) The transmission energy of (A) is as follows:
Figure RE-FDA0003062872960000061
Figure RE-FDA0003062872960000062
wherein, PTAnd PRWork of transmission respectively for nodesRate and received power, thus ε (ω)i) The activity energy consumption of (2) is as follows:
Figure RE-FDA0003062872960000063
forwarding node
Figure RE-FDA0003062872960000064
Includes only the transmission energy consumption, as in equation (18):
Figure RE-FDA0003062872960000065
from the equations (9) and (17), the mapping node ε (ω)i) The total energy consumption of (2) is as follows:
Figure 3
obtained by the equations (13) and (18), forwarding node
Figure RE-FDA0003062872960000072
The total energy consumption of (a) is as follows:
Figure 1
the total energy consumption of the entire mist network is shown as equation (21):
Figure RE-FDA0003062872960000074
let the maximum energy contained in the whole fog network be EmaxThen, in a certain task G, the energy consumption generated by the network needs to be less than or equal to the maximum energy consumption, as shown in equation (22):
Figure RE-FDA0003062872960000075
(24) constructing an optimization model of the mapping rule from the directed acyclic graph G in the DAG form to the undirected connected graph U, and based on the steps (21) and (23), providing the mapping rule from the directed acyclic graph G in the DAG form to the undirected connected graph U, optimizing the model, and establishing a binary optimization problem:
define 3. subtask node ωpAnd fog network node vqThe mapping relationship of (1) is as follows: when in use
Figure RE-FDA0003062872960000081
When it is, i.e. ωpIs mapped as vq(ii) a When in use
Figure RE-FDA0003062872960000082
Time, omegapWill not be mapped as vqThen, then
Figure RE-FDA0003062872960000083
Satisfies formula (23):
Figure RE-FDA0003062872960000084
based on the definition 3, ωpTo vqCan be constructed as a mapping matrix X of l X t, as in equation (24):
Figure RE-FDA0003062872960000085
the subtask ω represented by equation (5)iCan be expressed by equation (25):
Figure RE-FDA0003062872960000086
the latency of task G may be expressed as a function of X, as in equation (26):
T(G)＝F(X) (26)
the mapping node ε (ω) represented by formula (19)i) Can be expressed by the formula (27):
Figure RE-FDA0003062872960000087
the forwarding node represented by the formula (20)
Figure RE-FDA0003062872960000088
Can be expressed by equation (28):
Figure RE-FDA0003062872960000091
then, the optimal mapping relationship from the directed acyclic graph G to the undirected connected graph U under the energy consumption constraint is modeled as follows:
X＝arg min(F(X))
Figure RE-FDA0003062872960000092
4. the wireless sensor low-power-consumption low-delay path type collaborative computing method as claimed in claim 1, wherein the BPSO algorithm in step (3) is mainly used for optimizing a constraint problem of a discrete space, and limits the position of a particle to 0 or 1, and is applicable to a binary optimization problem proposed by formula (29):
when BPSO algorithm is adopted, particle swarm
Figure RE-FDA0003062872960000101
Moving within the search space I to find the best position,
Figure RE-FDA0003062872960000102
in NmaxIs the maximum value of the iteration times, M is the particle swarm size, n belongs to{1,2,…,NmaxThe iteration times are;
in the nth iteration, the position and velocity of the ith particle may be expressed as:
Figure RE-FDA0003062872960000103
Figure RE-FDA0003062872960000104
in the formula (30), Xn(i)∈I，
Figure RE-FDA0003062872960000105
In the formula (31), Vn(i)∈O，
Figure RE-FDA0003062872960000106
For the ith particle, in the nth iteration, the velocity update formula is as follows (32):
Figure RE-FDA0003062872960000107
in the formula (32), the compound represented by the formula (32),
Figure RE-FDA0003062872960000108
and
Figure RE-FDA0003062872960000109
respectively the local and global optimum positions of the particle, w is the inertial weight, gamma1And gamma2As an acceleration factor, beta1And beta2Is in the interval of [0,1 ]]Random numbers uniformly distributed therein;
the position update formula of the BPSO algorithm is as shown in formulas (33) to (34):
Figure RE-FDA00030628729600001010
Figure RE-FDA00030628729600001011
the fitness function of the algorithm is as follows (35):
f(X)＝T(G)＝F(X) (35)。
CN202110038750.8A 2021-01-12 2021-01-12 Wireless sensor low-power-consumption low-time-delay path type collaborative computing method Active CN112996073B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110038750.8A CN112996073B (en) 2021-01-12 2021-01-12 Wireless sensor low-power-consumption low-time-delay path type collaborative computing method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110038750.8A CN112996073B (en) 2021-01-12 2021-01-12 Wireless sensor low-power-consumption low-time-delay path type collaborative computing method

Publications (2)

Publication Number Publication Date
CN112996073A true CN112996073A (en) 2021-06-18
CN112996073B CN112996073B (en) 2022-10-28

Family

ID=76345323

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110038750.8A Active CN112996073B (en) 2021-01-12 2021-01-12 Wireless sensor low-power-consumption low-time-delay path type collaborative computing method

Country Status (1)

Country Link
CN (1) CN112996073B (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120134280A1 (en) * 2010-11-29 2012-05-31 Rosemount, Inc. Wireless sensor network access point and device rf spectrum analysis system and method
CN105848242A (en) * 2016-03-25 2016-08-10 黑龙江大学 Safety route optimization method based on trust perception in wireless sensor network
CN106413024A (en) * 2016-08-29 2017-02-15 华侨大学 Sensor cloud data transmission algorithm based on fog computing
CN106993320A (en) * 2017-03-22 2017-07-28 江苏科技大学 Wireless sensor network cooperation transmission method for routing based on many relay multi-hops
CN107343025A (en) * 2017-06-07 2017-11-10 西安电子科技大学 Time delay optimization method under the distributed satellites cloud and mist network architecture and power consumption constraint
CN109982283A (en) * 2019-02-15 2019-07-05 江苏商贸职业学院 A kind of industrial cloud and mist framework communication system for transmitting energy consumption towards expectation

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120134280A1 (en) * 2010-11-29 2012-05-31 Rosemount, Inc. Wireless sensor network access point and device rf spectrum analysis system and method
CN105848242A (en) * 2016-03-25 2016-08-10 黑龙江大学 Safety route optimization method based on trust perception in wireless sensor network
CN106413024A (en) * 2016-08-29 2017-02-15 华侨大学 Sensor cloud data transmission algorithm based on fog computing
CN106993320A (en) * 2017-03-22 2017-07-28 江苏科技大学 Wireless sensor network cooperation transmission method for routing based on many relay multi-hops
CN107343025A (en) * 2017-06-07 2017-11-10 西安电子科技大学 Time delay optimization method under the distributed satellites cloud and mist network architecture and power consumption constraint
CN109982283A (en) * 2019-02-15 2019-07-05 江苏商贸职业学院 A kind of industrial cloud and mist framework communication system for transmitting energy consumption towards expectation

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
LINA HUANG: "Software-defined QoS provisioning for fog computing advanced wireless sensor networks", 《2016 IEEE SENSORS》 *
何秀丽: "面向低时延的云雾网络及其负载均衡策略研究", 《中国优秀硕士学位论文全文数据库》 *
张夏童: "面向医疗大数据任务低时延需求的路径计算方案", 《西安交通大学学报》 *

Also Published As

Publication number Publication date
CN112996073B (en) 2022-10-28

Similar Documents

Publication Publication Date Title
Chen et al. When UAV swarm meets edge-cloud computing: The QoS perspective
Lei et al. Toward intelligent cooperation of UAV swarms: When machine learning meets digital twin
Tang et al. Survey on digital twin edge networks (DITEN) toward 6G
Cao et al. Swarm intelligence-based performance optimization for mobile wireless sensor networks: survey, challenges, and future directions
CN107343025B (en) Delay optimization method under distributed satellite cloud and mist network architecture and energy consumption constraint
CN111367657B (en) Computing resource collaborative cooperation method based on deep reinforcement learning
Lim et al. UAV-assisted communication efficient federated learning in the era of the artificial intelligence of things
CN113312177B (en) Wireless edge computing system and optimizing method based on federal learning
CN113238847B (en) Distribution and scheduling method based on distributed network environment and capable of distributing tasks
Yang et al. An adaptive immune algorithm for service-oriented agricultural Internet of Things
CN111885493B (en) Micro-cloud deployment method based on improved cuckoo search algorithm
CN114546608A (en) Task scheduling method based on edge calculation
CN110830292A (en) Medical big data-oriented cloud and mist mixed path determination method
CN114298178A (en) Multi-agent communication learning method
Sun et al. Flexedge: Dynamic task scheduling for a uav-based on-demand mobile edge server
Prakash et al. Energy and congestion-aware load balanced multi-path routing for wireless sensor networks in ambient environments
CN112996073B (en) Wireless sensor low-power-consumption low-time-delay path type collaborative computing method
CN116489708B (en) Meta universe oriented cloud edge end collaborative mobile edge computing task unloading method
Senthilkumar et al. Advertisement timeout driven bee's mating approach to maintain fair energy level in sensor networks
CN112579285A (en) Edge network-oriented distributed neural network collaborative optimization method
Qiao et al. A service function chain deployment scheme of the software defined satellite network
CN106209683A (en) Data transmission method based on data center's wide area network and system
CN115361453A (en) Load fair unloading and transferring method for edge service network
Panimalar et al. Data aggregation using squirrel search algorithm in wireless sensor networks
Gao et al. Fast Adaptive Task Offloading and Resource Allocation in Large-Scale MEC Systems via Multi-Agent Graph Reinforcement Learning

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