CN113709775B - Distributed cooperative computing migration method for satellite-ground cooperative communication system - Google Patents
Distributed cooperative computing migration method for satellite-ground cooperative communication system Download PDFInfo
- Publication number
- CN113709775B CN113709775B CN202111251939.1A CN202111251939A CN113709775B CN 113709775 B CN113709775 B CN 113709775B CN 202111251939 A CN202111251939 A CN 202111251939A CN 113709775 B CN113709775 B CN 113709775B
- Authority
- CN
- China
- Prior art keywords
- satellite
- calculation
- task
- base station
- computing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W24/00—Supervisory, monitoring or testing arrangements
- H04W24/02—Arrangements for optimising operational condition
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/08—Configuration management of networks or network elements
- H04L41/0803—Configuration setting
- H04L41/0823—Configuration setting characterised by the purposes of a change of settings, e.g. optimising configuration for enhancing reliability
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/08—Configuration management of networks or network elements
- H04L41/0803—Configuration setting
- H04L41/0823—Configuration setting characterised by the purposes of a change of settings, e.g. optimising configuration for enhancing reliability
- H04L41/083—Configuration setting characterised by the purposes of a change of settings, e.g. optimising configuration for enhancing reliability for increasing network speed
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
- H04L67/101—Server selection for load balancing based on network conditions
-
- 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
- H04W84/00—Network topologies
- H04W84/02—Hierarchically pre-organised networks, e.g. paging networks, cellular networks, WLAN [Wireless Local Area Network] or WLL [Wireless Local Loop]
- H04W84/04—Large scale networks; Deep hierarchical networks
- H04W84/06—Airborne or Satellite Networks
Abstract
The invention provides a distributed cooperative computing migration method of a satellite-ground cooperative communication system, which comprises the steps of collecting channel information and computing task information required by the satellite-ground cooperative communication system to finish computing migration; determining a system delay optimization problem based on a distributed collaborative computing migration mode; performing distributed solution on the system delay optimization problem to respectively obtain the optimal calculation migration strategy of each base station and each satellite; and based on the optimal calculation migration strategy, each base station and each satellite respectively perform edge calculation, and the result is returned to the user. The method can fully utilize a double-layer edge computing framework in the satellite-ground cooperative communication system, and avoids bidirectional propagation delay caused by a centralized decision process through a distributed migration mechanism, so that the system delay is remarkably reduced, the problem of decision delay time length caused by adopting a centralized migration strategy in the satellite-ground cooperative communication system is solved, and the transmission of user delay sensitive services is ensured.
Description
Technical Field
The invention relates to the technical field of resource allocation in wireless communication, in particular to a distributed cooperative computing migration method for a satellite-ground cooperative communication system.
Background
The development of satellite internet is rapid, the 6G wireless network is also proposed in the next generation mobile communication 6G white paper, the seamless connection among the ground, the satellite and the airborne network is realized, and the satellite-ground cooperative communication system is an important development direction of the next generation mobile communication. However, the user needs to go through the satellite-ground link to access the internet through the satellite, which causes a very large network delay. Taking a low-orbit satellite with a height of 1000km orbit as an example, the satellite-ground link delay reaches 13ms, while the satellite-ground link delay of an orbit satellite in a 20000km orbit reaches 270ms, which is far from reaching the delay indexes of 1ms for 5G communication and 0.1ms for 6G communication.
The rapid development of global smart phones has promoted the development of mobile terminals and edge computing, and edge computing systems have been called accordingly. By introducing the edge server, various processing works are completed on a local edge layer without uploading to the cloud, so that the processing efficiency is greatly improved, the user delay is reduced, and the network response speed is improved. By researching the satellite-ground cooperative system edge computing technology, the service is provided on the network edge closer to the terminal, the time delay caused by complex long-chain path transmission can be reduced, and the system service efficiency is improved.
Because the computing capacity of the edge server is limited, a migration strategy needs to be determined according to the computing task requirements of the user, and the system delay is minimized. However, due to the particularity of the satellite-ground cooperative system architecture, if a centralized computing migration method is adopted, the satellite needs to collect task information of all base stations and transmit the strategy to each base station through a satellite-ground link. This centralized optimization process will result in additional two-way propagation delay. In order to meet the requirements of diversified users in a satellite-ground cooperative system in the future and guarantee the quality of delay sensitive services, an innovative calculation migration method needs to be provided by using a cooperative calculation architecture in the satellite-ground cooperative communication system, so that the problem that a centralized migration strategy is adopted to decide the delay time in the satellite-ground cooperative communication system is solved.
Disclosure of Invention
The invention aims to provide a distributed cooperative computing migration method for a satellite-ground cooperative communication system, which is based on a double-layer edge computing architecture of the satellite-ground cooperative communication system, designs a distributed cooperative computing migration mechanism, provides edge computing service for users, solves the problem of delay time of decision by adopting a centralized migration strategy in the satellite-ground cooperative communication system, and ensures the transmission of delay sensitive services of the users.
In order to achieve the purpose, the invention provides the following technical scheme:
the application discloses a distributed collaborative computing migration method of a satellite-ground collaborative communication system, which comprises the following steps:
s1, collecting channel information and calculation task information required by the satellite-ground cooperative communication system to complete calculation migration; the satellite-ground cooperative communication system comprises a space section and a ground section; the space segment is composed of satellites, and the ground segment is composed of users and a base station; the user generates a computing task;
s2, determining a system delay optimization problem based on a distributed collaborative computing migration mode;
s3, carrying out distributed solution on the system delay optimization problem to respectively obtain the distributed calculation migration strategy of each base station and each satellite;
and S4, based on the distributed computing migration strategy, each base station and each satellite respectively perform edge computing, and the result is returned to the user.
Preferably, in step S1, the satellite covers several base stations, and the satellite is connected to the base stations in its coverage area; each base station is covered with a plurality of users, and the base station is connected with the users in the coverage range; and edge servers which can be used for computing services are arranged in the base station and the satellite.
Preferably, in step S1, the channel information is uplink average transmission channel information from each user to the base station; the calculation task information comprises the input data amount of each calculation task and the CPU calculation amount required by each bit of data of each calculation task.
Preferably, the step S2 specifically includes the following sub-steps:
s21, calculating the average uplink transmission rate from each user to the base station, and calculating the average uplink transmission delay based on the average uplink transmission rate; calculating the calculation delay of the calculation task generated by each user in the base station, and calculating the satellite calculation delay estimated by the calculation task of each user in the base station; calculating the calculation delay of each user calculation task in the satellite;
s22, summing the average uplink transmission delay of the users within the coverage range of each base station, and minimizing the sum of the average uplink transmission delay under the constraint condition of the length of the time frame to obtain a sub-problem of time slot allocation;
s23, summing the total task calculation delays of the users in the coverage area of each base station, and minimizing the sum of the total task calculation delays of all the users in the coverage area under the constraint of task decomposition and the constraint of total calculation capacity of the base station to obtain the sub-problems of task decomposition and base station calculation resource allocation; the task total calculation delay is the larger of the base station calculation delay and the satellite calculation delay of the calculation task;
s24, summing satellite calculation time delays of calculation tasks transferred to the satellites, and minimizing the sum of the satellite calculation time delays under the constraint of the maximum allocable satellite calculation resources of the tasks and the constraint of the total satellite calculation capacity to obtain satellite calculation resource allocation sub-problems; the task maximum allocable satellite computing resource constraint is that the satellite computing time delay of the computing task is equal to the satellite computing resource allocated by the computing time delay of the computing task base station;
s25, minimizing the total system delay under the system constraint to obtain the system delay optimization problem; the system constraint comprises a time frame length constraint of each base station, a calculation task decomposition constraint of each calculation task, a total calculation capacity constraint of each base station and a total calculation capacity constraint of a satellite, and the total system delay is the sum of the average uplink transmission delay of all users in the system and the total calculation delay of the tasks.
Preferably, the step S3 specifically includes the following sub-steps:
s31, solving the sub-problem of time slot allocation in the step S22 to obtain a time slot allocation strategy of the base station;
s32, solving the sub-problems of the calculation task decomposition and the base station calculation resource allocation in the step S23 to obtain a user calculation task decomposition strategy and a base station calculation resource allocation strategy in the coverage area of the base station;
s33, solving the satellite computing resource allocation subproblem in the step S24 to obtain a satellite computing resource allocation strategy;
preferably, the step S32 specifically includes the following iterative process:
s321, estimating total calculation data volume transferred to the satellite based on historical information; the historical information is total data size information transferred to the satellite in the historical calculation transfer process;
s322, initializing a base station to calculate a resource allocation strategy and an upper limit of iteration times;
s323, updating the base station computing resource allocation strategy based on the resource updating strategy; the resource updating strategy is a strategy for reducing the sum of the task total calculation delays of all users in the coverage area of the base station;
s324, estimating satellite computing resources allocated to each computing task based on the updated base station computing resource allocation strategy;
s325, calculating a task decomposition strategy according to the updated base station calculation resource allocation strategy and the estimated satellite calculation resource allocation strategy;
s326, calculating the sum of the total calculation delay of the tasks of all users in the coverage range of the base station according to the updated base station calculation resource allocation strategy, the estimated satellite calculation resource allocation strategy and the calculated task decomposition strategy;
s327, comparing the sum of the task total calculation delays of all updated users with the sum of the task total calculation delays of all updated users, and setting a base station calculation resource allocation strategy as a delay reduction strategy;
s328, judging whether the iteration times reach an upper limit;
s329, if the upper limit is reached, finishing iteration, setting the finally obtained base station computing resource allocation strategy as the base station computing resource allocation strategy obtained in the step S327, and setting the finally obtained decomposition strategy as the task decomposition strategy calculated in the step S325 based on the base station computing resource allocation strategy;
and S3210, if the upper limit is not reached, returning to the step S323 to repeat the iteration process until the iteration is finished.
Preferably, the step S33 includes the following iterative process:
s331, initializing a current computing task set into all task sets migrated to a satellite, initializing current satellite computing capacity into satellite total computing capacity, and initializing dual satellite computing resource allocation problems into satellite computing resource allocation sub-problems;
s332, removing the constraint of satellite computing resources with the maximum task assignable in the dual satellite computing resource allocation problem to obtain a loose satellite computing resource allocation problem;
s333, calculating a solution of the calculation resource allocation problem of the loose satellite to obtain satellite calculation resources allocated to each calculation task in the calculation resource allocation problem of the loose satellite;
s334, judging whether a task with the calculation resource allocation exceeding the maximum task allocable satellite calculation resource constraint exists in the solution of the calculation resource allocation problem of the loose satellite;
s335, if the task with the distributed computing resources exceeding the constraint of the maximum distributable satellite computing resources of the task exists, enabling the excess set to be the task set with the distributed computing resources exceeding the constraint of the maximum distributable satellite computing resources of the task, and setting the solution of the satellite computing resources distributed by the task in the excess set in the satellite computing resource distribution sub-problem as the maximum distributable satellite computing resources;
s3351, updating the current calculation task set to subtract the excess set from the current calculation task set, and updating the current satellite calculation capacity to subtract the calculation capacity distributed by the tasks in the excess set from the current satellite calculation capacity;
s3352, updating the dual satellite computing resource allocation problem based on the updated current computing task set and the current satellite computing capacity, returning to the step S332, and repeatedly executing the iteration process until the iteration is finished;
and S336, if the task with the distributed computing resources exceeding the constraint of the maximum distributable satellite computing resources of the task does not exist, setting the solution of the task in the satellite computing resource distribution sub-problem in the current computing task set as the solution of the problem of loose satellite computing resource distribution obtained in the step S333, and ending the iteration.
Preferably, the step S4 specifically includes the following operations: each base station migrates part of the calculation tasks to the satellite according to the calculation task decomposition strategy obtained by solving, performs edge calculation according to the base station calculation resource allocation strategy obtained by solving, and returns the calculation result to the user; the satellite carries out edge calculation based on the obtained satellite calculation resource allocation strategy, the calculation result is returned to the base station, and the base station further returns the result to the user.
The invention has the beneficial effects that: compared with the prior art, the distributed collaborative computing migration method of the satellite-ground collaborative communication system provided by the invention firstly collects the channel information and the computing task information required by the satellite-ground collaborative communication system to complete computing migration; then, determining a system delay optimization problem based on a distributed collaborative computing migration mode; then, carrying out distributed solution on the system delay optimization problem to respectively obtain a distributed computing migration strategy of each base station and each satellite; finally, based on the distributed computation migration strategy, each base station and each satellite respectively carry out edge computation, and the result is returned to the user; the method can fully utilize a double-layer edge computing framework in the satellite-ground cooperative communication system, each base station and each satellite can perform decision making and computing migration in a distributed mode through a distributed migration mechanism without waiting for a centralized strategy of the satellite, and bidirectional propagation delay caused by the centralized decision making process is avoided, so that system delay is remarkably reduced, the problem of delay time of decision making by adopting the centralized migration strategy in the satellite-ground cooperative communication system is solved, and transmission of delay sensitive services of users is guaranteed.
The features and advantages of the present invention will be described in detail by embodiments in conjunction with the accompanying drawings.
Drawings
Fig. 1 is a flowchart of a distributed collaborative computing migration method of a satellite-ground collaborative communication system according to an embodiment of the present invention;
fig. 2 is a schematic diagram of a satellite-ground cooperative communication system according to an embodiment of the present invention;
fig. 3 is a flowchart of collecting channel information and computation task information required by the satellite-ground cooperative communication system to complete computation migration according to the embodiment of the present invention;
FIG. 4 is a flow chart of a system delay optimization problem determination based on a distributed collaborative computing migration approach as provided by an embodiment of the present invention;
fig. 5 is a flowchart for performing distributed solution on the system delay optimization problem to obtain distributed computation migration strategies for each base station and each satellite according to the embodiment of the present invention;
fig. 6 is a flowchart for solving the sub-problems of the computation task decomposition and the base station computation resource allocation for each base station to obtain a user computation task decomposition strategy and a base station computation resource allocation strategy within the coverage area of the base station according to the embodiment of the present invention;
fig. 7 is a flowchart for solving the satellite computing resource allocation sub-problem to obtain a satellite computing resource allocation policy according to the embodiment of the present invention;
fig. 8 is a schematic diagram illustrating comparison of performances of a distributed collaborative computing migration method of a satellite-ground collaborative communication system according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood, however, that the description herein of specific embodiments is only illustrative of the invention and is not intended to limit the scope of the invention. Moreover, in the following description, descriptions of well-known structures and techniques are omitted so as to not unnecessarily obscure the concepts of the present invention.
The first embodiment is as follows:
a method for migrating distributed collaborative computing in a satellite-ground collaborative communication system, referring to fig. 1, the method comprising:
s101, collecting channel information and calculation task information required by a satellite-ground cooperative communication system to finish calculation migration;
in the embodiment of the invention, the execution main body of the method is the base station, and each base station only collects the channel information and the calculation task information of the users in the coverage area.
The process of collecting channel information and calculation task information required for the satellite-ground cooperative communication system to complete calculation migration is described in detail below, and is not described herein again.
S102, determining a system delay optimization problem based on a distributed collaborative computing migration mode;
the distributed collaborative computing migration mode means that each base station independently performs computing migration on computing tasks of users in the coverage area of the base station without waiting for a centering strategy of a satellite, and the satellite independently performs computing migration on computing tasks migrated to the satellite. The system delay optimization problem comprises channel information and calculation task information, and the system delay optimization problem represents that the total delay from generation of calculation tasks to acquisition of calculation results of all users is minimized under the condition that system resource constraints are met. The process is described in detail below and will not be described herein.
S103, carrying out distributed solution on the system delay optimization problem to respectively obtain a distributed computing migration strategy of each base station and each satellite;
the distributed solving means that each base station independently solves the base station distributed computing migration strategy, and the satellite independently solves the satellite distributed computing migration strategy. The process is described in detail below and will not be described herein.
And S104, based on the distributed computing migration strategy, each base station and each satellite respectively perform edge computing, and return the result to the user.
Due to the particularity of the satellite-ground cooperative system architecture, if a centralized computing migration method is adopted, the satellite needs to collect task information of all base stations and transmits a strategy to each base station through a satellite-ground link, which causes additional bidirectional propagation delay. Compared with the existing centralized computing migration method, in the distributed collaborative computing migration method of the satellite-ground collaborative communication system in the embodiment of the invention, channel information and computing task information required by the satellite-ground collaborative communication system for completing computing migration are collected firstly; then, determining a system delay optimization problem based on a distributed collaborative computing migration mode; then, carrying out distributed solution on the system delay optimization problem to respectively obtain a distributed computing migration strategy of each base station and each satellite; finally, based on the distributed computation migration strategy, each base station and each satellite respectively carry out edge computation, and the result is returned to the user; the method can fully utilize a double-layer edge computing framework in the satellite-ground cooperative communication system, each base station and each satellite can perform decision making and computing migration in a distributed mode through a distributed migration mechanism without waiting for a centralized strategy of the satellite, and bidirectional propagation delay caused by the centralized decision making process is avoided, so that system delay is remarkably reduced, the problem of delay time of decision making by adopting the centralized migration strategy in the satellite-ground cooperative communication system is solved, and transmission of delay sensitive services of users is guaranteed.
The foregoing briefly introduces a method for migrating distributed collaborative computing in a satellite-ground collaborative communication system, and the following describes the specific details involved in the method in detail.
In an alternative embodiment, referring to fig. 2, a satellite-to-ground cooperative communication system includes:
the satellite-ground cooperative communication system comprises a space section and a ground section;
the space section is composed of satellites, the satellites provide return transmission for a ground base station lacking optical fiber coverage, the satellites can be low-orbit satellites, medium-orbit satellites and high-orbit satellites according to actual network composition, and transmission delay and service capacities of the satellites in different orbits are different;
the satellite is connected with the base station in the coverage range of the satellite, an edge server exists in the satellite and can provide edge computing service, and the satellite provides the edge computing service for the base station in the coverage range of the satellite;
the ground segment is composed of users and a base station;
the total number of base stations in the system isWithin the service coverage of each base stationThe user is only connected with the subordinate base station, the subordinate base station means that the user is positioned in the coverage range of the base station, and the user obtains edge calculation service through the base station;
the base station is connected with users in the coverage range and the satellite, an edge server exists in the base station and can provide edge computing service, the base station provides edge computing service for the users in the coverage range and can transfer the computing task part of the users to the satellite for computing;
to any base stationAny user in the coverage areaThe user only generates one calculation task at the same time, and the calculation process is as follows:
Base stationUser will beThe calculation task of (2) is divided into two parts, one part carries out edge calculation at the base station, and the other part is transmitted to the satellite;
user' sAfter the part of the computing task is transmitted to the satellite, the satellite performs computing on the part of the computing task.
Optionally, referring to fig. 3, collecting channel information and computation task information required by the satellite-ground cooperative communication system to complete computation migration includes:
s201, collecting uplink average transmission channel information from each user to a base station;
the average transmission channel information represents the average transmission channel information in the process of completing the transmission of calculation tasks by the user, the average value can be obtained based on the large-scale fading of the channel and the historical channel information, and the user can use the average transmission channel information to calculate the channel quality of the channel qualityTo the base stationIs expressed as。
S202, collecting input data quantity of each calculation task and CPU calculation quantity required by each bit of data of each calculation task;
in the system, each user has a computing task that requires computing. For base stationUsers in coverageThe computing task is represented asWhereinIn order to calculate the amount of input data for a task,is the amount of CPU computation required to compute each bit of data for a task, so that the total amount of computation required for each task is。
Optionally, referring to fig. 4, determining a system delay optimization problem based on a distributed collaborative computing migration method includes:
s301, for each user, calculating the average uplink transmission delay from the user to a base station, calculating the calculation delay of a user calculation task at the base station, calculating the satellite calculation delay estimated by the user calculation task at the base station, and calculating the calculation delay of the user calculation task at the satellite;
for each user, communicating with the base station based on time division multiple access, migrating the computation task to the base station. Order toIn order to be able to transmit the bandwidth,is the transmission power of the user for which,is the noise power spectral density, so the average uplink transmission rate from the user to the base station is:
order toFor each time frame length, to usersHas a slot length ofAnd the time slot allocation strategy is kept unchanged in the transmission process, so that the average uplink transmission delay from the user to the base station is as follows:
in addition, the return data volume of the calculation result is far smaller than the input data volume of the calculation task, so that the downlink transmission delay for retrieving the calculation result from the base station is ignored.
Order toRepresenting computational tasksThe proportion of data calculated at the base station,indicating allocated base station computing resources, in computing tasksAfter migrating to the base station, computing tasksThe calculated delay at the base station is:
if it is notThe task will be further migrated to the satellite, orderAnd (3) representing the estimated and distributed satellite computing resources, wherein the estimated satellite computing delay of the task at the base station is as follows:
order toRepresenting the actually allocated satellite computing resources, the satellite computing delay of the task is:
based on a large antenna and Ku/Ka frequency band broadband transmission, the uplink transmission from a base station to a satellite can realize higher transmission rate from hundreds of MB/s to GB/s. On the other hand, the amount of data input by the user's computing task is typically on the order of tens of KB to hundreds of KB, much smaller than the base station uplink transmission rate, so the transmission delay from the base station to the satellite, and the transmission delay from the satellite back to the base station, can be neglected. Due to the long time delay of the satellite-ground link, when the calculation task is migrated to the satellite, the bidirectional satellite-ground propagation time delay caused by migration of the calculation task to the satellite still needs to be considered, which is expressed as:
S302, summing the average uplink transmission delay of the users in the coverage area of each base station, and minimizing the sum of the average uplink transmission delay under the constraint condition of the length of a time frame to obtain a sub-problem of time slot allocation;
s303, minimizing the sum of the total task calculation delays of all users in the coverage area of each base station under the conditions of calculation task decomposition constraint and base station total calculation capacity constraint to obtain the problem of calculation task decomposition and base station calculation resource allocation sub-units;
order toMaximum calculation capacity allocable for each base station, for a base stationThe sub-problems of the calculation task decomposition and the base station calculation resource allocation are as follows:
s304, for the satellite, minimizing the sum of satellite calculation time delays under the constraint of the maximum allocable satellite calculation resource of the task and the constraint of the total calculation capacity of the satellite to obtain a satellite calculation resource allocation sub-problem;
order toFor the maximum computational capacity that the satellite can allocate,for the set of computing tasks migrated to the satellite, the satellite computing resource allocation sub-problem is:
S305, minimizing the total system delay under the system constraint to obtain a system delay optimization problem;
the system delay optimization problem is as follows:
optionally, referring to fig. 5, performing distributed solution on the system delay optimization problem to obtain the distributed computing migration policy of each base station and each satellite respectively includes:
s401, solving a time slot allocation subproblem for each base station to obtain a time slot allocation strategy of the base station;
s402, solving the sub-problems of calculation task decomposition and base station calculation resource allocation for each base station to obtain a user calculation task decomposition strategy and a base station calculation resource allocation strategy in the coverage area of the base station;
after each base station receives the calculation tasks of the users, the sub-problems of calculation task decomposition and base station calculation resource allocation are solved in a distributed mode, under the conditions of calculation task decomposition constraint and base station total calculation capacity constraint, a user calculation task decomposition strategy and a base station calculation resource allocation strategy in the coverage range of the base station are obtained, and the sum of the total calculation delay of the tasks of all the users in the coverage range is minimized. The process is described in detail below and will not be described herein.
S403, solving the sub-problem of satellite computing resource allocation to obtain a satellite computing resource allocation strategy;
after the satellite receives the calculation task of the migrated satellite, the satellite calculation resource allocation sub-problem is solved in a distributed mode, a satellite calculation resource allocation strategy is obtained under the constraint of the satellite calculation resource with the maximum task allocable and the constraint of the total satellite calculation capacity, and the sum of satellite calculation delays of the calculation task migrated to the satellite is minimized. The process is described in detail below and will not be described herein.
Optionally, referring to fig. 6, for each base station, solving the sub-problem of computing task decomposition and base station computing resource allocation to obtain a user computing task decomposition policy and a base station computing resource allocation policy within a coverage area of the base station includes:
s501, estimating total calculation data volume transferred to a satellite based on historical information;
the historical information is total data volume information transferred to the satellite in the historical calculation transfer process, the information is returned to each base station by the satellite in each calculation transfer process, and the historical information calculation formula is as follows:
the pre-estimated formula is as follows:
whereinIs as followsPeriodic pairIs determined by the estimated value of (c),is as followsPeriod of timeThe actual value of (a) is,is a smoothing parameter.
S502, initializing a base station to calculate a resource allocation strategy and an upper limit of iteration times;
Denotes a base stationA possible computing resource allocation policy of, whereinIndicating a base stationAssigning to tasksThe computing resources of (1). The velocity vector of each particle isAnd represents the moving tendency of each particle. During the search, each particle will record the local best position that the particle has experiencedAt the same time, useThe global best position experienced by all particles is recorded.
S503, updating the base station computing resource allocation strategy based on the resource updating strategy;
in each iteration, the position and velocity of each particle is updated as follows
WhereinIs the number of iterations that are to be performed,is the weight of the inertia, and,is the acceleration factor, and is,are independent random variables.
S504, estimating satellite computing resources allocated to each computing task based on the updated base station computing resource allocation strategy;
for each particle, the location vector corresponds to the base station computing resource allocation policy,based on the base station computing resource allocation strategy, judging tasks according to the following strategyWhether to migrate to satellite:
if it isTask ofIt will not migrate to the satellite and all calculations will be performed at the base station.
If it isTask ofThe mobile station is moved to the satellite, and the calculation is carried out at the base station and the satellite simultaneously, and is distributed to the taskThe estimated satellite resources are
S505, calculating a task decomposition strategy according to the updated base station calculation resource allocation strategy and the estimated satellite calculation resource allocation strategy;
s506, calculating the sum of the total calculation delay of the tasks of all users in the coverage area of the base station according to the updated base station calculation resource allocation strategy, the estimated satellite calculation resource allocation strategy and the calculated task decomposition strategy;
order toIndicating particleThe base station obtained by calculation under the corresponding base station calculation resource allocation strategyThe sum of the calculated delays for all users' tasks in the coverage area is expressed as:
s507, comparing the sum of the task total calculation delays of all the updated users with the sum of the task total calculation delays of all the updated users, and setting a base station calculation resource allocation strategy as a delay reduction strategy;
local best position update of each particle to
Global best location update as
And the base station computing resource allocation strategy is a computing resource allocation strategy corresponding to the global best position.
S508, judging whether the iteration number reaches an upper limit;
S509, if the upper limit is reached, iteration is finished, a base station computing resource allocation strategy and a task decomposition strategy are obtained;
the base station calculates the resource allocation strategy as the global best positionAnd (4) corresponding computing resource allocation strategies.
The calculation task decomposition policy is calculated based on step S505.
S510, if the upper limit is not reached, repeatedly executing an iteration process until the iteration is finished;
returning to step S503, the iteration is repeatedly performed until the iteration is finished.
Optionally, referring to fig. 7, solving the satellite computing resource allocation sub-problem to obtain the satellite computing resource allocation policy includes:
s601, initializing a current computing task set into all task sets migrated to a satellite, initializing current satellite computing capacity into satellite total computing capacity, and initializing a dual satellite computing resource allocation problem into a satellite computing resource allocation sub-problem;
The problem of initializing dual satellite resource allocation is as follows:
s602, removing the constraint of satellite computing resources with the maximum task assignable in the dual satellite computing resource allocation problem to obtain a loose satellite computing resource allocation problem;
the problem of the allocation of the calculation resources of the relaxation satellite is as follows:
s603, calculating a solution of the calculation resource distribution problem of the loose satellite to obtain satellite calculation resources distributed to each calculation task in the calculation resource distribution problem of the loose satellite;
the satellite computing resources allocated to each computing task in the relaxed satellite computing resource allocation problem are:
s604, judging whether a task with the computing resource allocation exceeding the task maximum allocable satellite computing resource constraint exists in the solution of the loose satellite computing resource allocation problem;
S605, if the resource allocation problem exists, the excess set is a task set of which the allocation computing resource exceeds the constraint of the maximum allocable satellite computing resource of the task, and the solution of the satellite computing resource allocated by the task in the excess set is set as the maximum allocable satellite computing resource in the satellite computing resource allocation problem;
order toTo satisfyTask set, order set ofSolution of the satellite computing resource allocation sub-problem of the medium task。
S606, updating the current calculation task set to subtract the excess set from the current calculation task set, and updating the current satellite calculation capacity to subtract the calculation capacity distributed by the tasks in the excess set from the current satellite calculation capacity;
S607, updating the dual satellite computing resource allocation problem based on the updated current computing task set and the current satellite computing capacity, and repeatedly executing the iteration process until the iteration is finished;
the problem of updating dual satellite resource allocation is as follows:
returning to step S602, the iteration process is repeatedly executed until the iteration is finished.
S608, if the task does not exist, setting the solution of the task in the current computing task set in the satellite computing resource allocation sub-problem as the solution of the satellite computing resource allocation problem, and ending iteration;
The invention provides a distributed collaborative computing migration method for a satellite-ground collaborative communication system, which is based on a double-layer edge computing architecture of the satellite-ground collaborative communication system, provides edge computing service for a user and returns a result to the user. The invention carries out distributed solution on the system delay optimization problem, and respectively obtains the distributed calculation migration strategy of each base station and each satellite by carrying out equivalent decomposition and decoupling on the original calculation migration problem, and finally obtains the calculation migration strategy with the minimum total system delay.
A schematic diagram of the satellite-ground cooperative communication system of the present invention is shown in FIG. 2, whereinThe total number of base stations is 10, each serving 5 users in a coverage area of 250 m. For each task, the amount of input data is [100, 500 ]]Randomly distributed in Kb range, the calculation amount required by each bit is 1000 CPU cycles, and the total calculation capacity of the base station iscycle/s, total calculated satellite capacity ofcycle/s. Referring to fig. 8 (comparing the proposed distributed computing migration policy with the centralized migration policy and the average migration policy), it can be seen that the proposed distributed computing migration policy can effectively reduce the satellite-ground collaborative communication system computing delay, which can reduce the total system delay by 20% compared to the centralized computing migration policy.
The computer program product of the distributed collaborative computing migration method for a satellite-ground collaborative communication system according to the embodiment of the present invention includes a computer-readable storage medium storing a program code, where instructions included in the program code may be used to execute the method described in the foregoing method embodiment, and specific implementation may refer to the method embodiment, which is not described herein again.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the system and the apparatus described above may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents or improvements made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (4)
1. A distributed cooperative computing migration method for a satellite-ground cooperative communication system is characterized by comprising the following steps:
s1, collecting channel information and calculation task information required by the satellite-ground cooperative communication system to complete calculation migration; the satellite-ground cooperative communication system comprises a space section and a ground section; the space segment is composed of satellites, and the ground segment is composed of users and a base station; the user generates a computing task;
s2, determining a system delay optimization problem based on a distributed collaborative computing migration mode;
s21, calculating the average uplink transmission rate from each user to the base station, and calculating the average uplink transmission delay based on the average uplink transmission rate; calculating the calculation delay of the calculation task generated by each user in the base station, and calculating the satellite calculation delay estimated by the calculation task of each user in the base station; calculating the calculation delay of each user calculation task in the satellite;
s22, summing the average uplink transmission delay of the users within the coverage range of each base station, and minimizing the sum of the average uplink transmission delay under the constraint condition of the length of the time frame to obtain a sub-problem of time slot allocation;
s23, summing the total task calculation delays of the users in the coverage area of each base station, and minimizing the sum of the total task calculation delays of all the users in the coverage area under the constraint of task decomposition and the constraint of total calculation capacity of the base station to obtain the sub-problems of task decomposition and base station calculation resource allocation; the task total calculation delay is the larger of the base station calculation delay and the satellite calculation delay of the calculation task;
s24, summing satellite calculation time delays of calculation tasks transferred to the satellites, and minimizing the sum of the satellite calculation time delays under the constraint of the maximum allocable satellite calculation resources of the tasks and the constraint of the total satellite calculation capacity to obtain satellite calculation resource allocation sub-problems; the task maximum allocable satellite computing resource constraint is that the satellite computing time delay of the computing task is equal to the satellite computing resource allocated by the computing time delay of the computing task base station;
s25, minimizing the total system delay under the system constraint to obtain the system delay optimization problem; the system constraint comprises a time frame length constraint of each base station, a calculation task decomposition constraint of each calculation task, a total calculation capacity constraint of each base station and a total calculation capacity constraint of a satellite, and the total system delay is the sum of the average uplink transmission delay of all users in the system and the total calculation delay of the tasks;
s3, carrying out distributed solution on the system delay optimization problem to respectively obtain the distributed calculation migration strategy of each base station and each satellite;
s31, solving the sub-problem of time slot allocation in the step S22 to obtain a time slot allocation strategy of the base station;
s32, solving the sub-problems of the calculation task decomposition and the base station calculation resource allocation in the step S23 to obtain a user calculation task decomposition strategy and a base station calculation resource allocation strategy in the coverage area of the base station;
s321, estimating total calculation data volume transferred to the satellite based on historical information; the historical information is total data size information transferred to the satellite in the historical calculation transfer process;
s322, initializing a base station to calculate a resource allocation strategy and an upper limit of iteration times;
s323, updating the base station computing resource allocation strategy based on the resource updating strategy; the resource updating strategy is a strategy for reducing the sum of the task total calculation delays of all users in the coverage area of the base station;
s324, estimating satellite computing resources allocated to each computing task based on the updated base station computing resource allocation strategy;
s325, calculating a task decomposition strategy according to the updated base station calculation resource allocation strategy and the estimated satellite calculation resource allocation strategy;
s326, calculating the sum of the total calculation delay of the tasks of all users in the coverage range of the base station according to the updated base station calculation resource allocation strategy, the estimated satellite calculation resource allocation strategy and the calculated task decomposition strategy;
s327, comparing the sum of the task total calculation delays of all updated users with the sum of the task total calculation delays of all updated users, and setting a base station calculation resource allocation strategy as a delay reduction strategy;
s328, judging whether the iteration times reach an upper limit;
s329, if the upper limit is reached, finishing iteration, setting the finally obtained base station computing resource allocation strategy as the base station computing resource allocation strategy obtained in the step S327, and setting the finally obtained decomposition strategy as the task decomposition strategy calculated in the step S325 based on the base station computing resource allocation strategy;
s3210, if the upper limit is not reached, returning to the step S323 to repeat the iteration process until the iteration is finished;
s33, solving the satellite computing resource allocation subproblem in the step S24 to obtain a satellite computing resource allocation strategy;
s331, initializing a current computing task set into all task sets migrated to a satellite, initializing current satellite computing capacity into satellite total computing capacity, and initializing dual satellite computing resource allocation problems into satellite computing resource allocation sub-problems;
s332, removing the constraint of satellite computing resources with the maximum task assignable in the dual satellite computing resource allocation problem to obtain a loose satellite computing resource allocation problem;
s333, calculating a solution of the calculation resource allocation problem of the loose satellite to obtain satellite calculation resources allocated to each calculation task in the calculation resource allocation problem of the loose satellite;
s334, judging whether a task with the calculation resource allocation exceeding the maximum task allocable satellite calculation resource constraint exists in the solution of the calculation resource allocation problem of the loose satellite;
s335, if the task with the distributed computing resources exceeding the constraint of the maximum distributable satellite computing resources of the task exists, enabling the excess set to be the task set with the distributed computing resources exceeding the constraint of the maximum distributable satellite computing resources of the task, and setting the solution of the satellite computing resources distributed by the task in the excess set in the satellite computing resource distribution sub-problem as the maximum distributable satellite computing resources;
s3351, updating the current calculation task set to subtract the excess set from the current calculation task set, and updating the current satellite calculation capacity to subtract the calculation capacity distributed by the tasks in the excess set from the current satellite calculation capacity;
s3352, updating the dual satellite computing resource allocation problem based on the updated current computing task set and the current satellite computing capacity, returning to the step S332, and repeatedly executing the iteration process until the iteration is finished;
s336, if the task with the distributed computing resources exceeding the constraint of the maximum distributable satellite computing resources of the task does not exist, setting the solution of the task in the satellite computing resource distribution sub-problem in the current computing task set as the solution of the problem of loose satellite computing resource distribution obtained in the step S333, and finishing the iteration;
and S4, based on the distributed computing migration strategy, each base station and each satellite respectively perform edge computing, and the result is returned to the user.
2. The method for distributed collaborative computing migration in a satellite-to-ground collaborative communication system according to claim 1, wherein in step S1, the satellite is covered with a plurality of base stations, and the satellite is connected to the base stations within the coverage area of the satellite; each base station is covered with a plurality of users, and the base station is connected with the users in the coverage range; and edge servers which can be used for computing services are arranged in the base station and the satellite.
3. The method for distributed cooperative computing migration in a satellite-to-ground cooperative communication system according to claim 1, wherein in step S1, the channel information is an uplink average transmission channel information from each user to a base station; the calculation task information comprises the input data amount of each calculation task and the CPU calculation amount required by each bit of data of each calculation task.
4. The method for migrating distributed collaborative computing in a satellite-ground collaborative communication system according to claim 1, wherein the step S4 specifically includes the following operations: each base station migrates part of the calculation tasks to the satellite according to the calculation task decomposition strategy obtained by solving, performs edge calculation according to the base station calculation resource allocation strategy obtained by solving, and returns the calculation result to the user; the satellite carries out edge calculation based on the obtained satellite calculation resource allocation strategy, the calculation result is returned to the base station, and the base station further returns the result to the user.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111251939.1A CN113709775B (en) | 2021-10-27 | 2021-10-27 | Distributed cooperative computing migration method for satellite-ground cooperative communication system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111251939.1A CN113709775B (en) | 2021-10-27 | 2021-10-27 | Distributed cooperative computing migration method for satellite-ground cooperative communication system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113709775A CN113709775A (en) | 2021-11-26 |
CN113709775B true CN113709775B (en) | 2022-01-07 |
Family
ID=78646993
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111251939.1A Active CN113709775B (en) | 2021-10-27 | 2021-10-27 | Distributed cooperative computing migration method for satellite-ground cooperative communication system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113709775B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114520990B (en) * | 2022-01-24 | 2023-05-26 | 重庆邮电大学 | Mobile edge computing task unloading method based on star-ground cloud cooperation |
CN114827152B (en) * | 2022-07-01 | 2022-09-27 | 之江实验室 | Low-delay cloud edge-side cooperative computing method and device for satellite-ground cooperative network |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110647391B (en) * | 2019-09-27 | 2022-04-12 | 北京邮电大学 | Edge computing method and system for satellite-ground cooperative network |
CN111611069B (en) * | 2020-04-01 | 2023-11-07 | 西南电子技术研究所(中国电子科技集团公司第十研究所) | Multi-type task migration method among multiple data centers |
CN112929427A (en) * | 2021-01-27 | 2021-06-08 | 北京邮电大学 | Service node determination method and device for low-earth-orbit satellite edge calculation |
-
2021
- 2021-10-27 CN CN202111251939.1A patent/CN113709775B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN113709775A (en) | 2021-11-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113709775B (en) | Distributed cooperative computing migration method for satellite-ground cooperative communication system | |
CN111586696B (en) | Resource allocation and unloading decision method based on multi-agent architecture reinforcement learning | |
CN112422644B (en) | Method and system for unloading computing tasks, electronic device and storage medium | |
Banoth et al. | Dynamic mobile charger scheduling with partial charging strategy for WSNs using deep-Q-networks | |
CN113543176A (en) | Unloading decision method of mobile edge computing system based on assistance of intelligent reflecting surface | |
Lyu et al. | Computation bits maximization in UAV-enabled mobile-edge computing system | |
CN111107566A (en) | Unloading method based on collaborative content caching in power Internet of things scene | |
CN116489708B (en) | Meta universe oriented cloud edge end collaborative mobile edge computing task unloading method | |
CN115175217A (en) | Resource allocation and task unloading optimization method based on multiple intelligent agents | |
Cui et al. | Latency Optimization for Hybrid GEO–LEO Satellite-Assisted IoT Networks | |
CN113821346B (en) | Edge computing unloading and resource management method based on deep reinforcement learning | |
CN114007231A (en) | Heterogeneous unmanned aerial vehicle data unloading method and device, electronic equipment and storage medium | |
CN112702443A (en) | Multi-satellite multi-level cache allocation method and device for satellite-ground cooperative communication system | |
Zhang et al. | Optimized look-ahead offloading decisions using deep reinforcement learning for battery constrained mobile iot devices | |
CN116193396A (en) | Slice-based collaborative task unloading method in air-space-ground integrated Internet of vehicles | |
CN116318371A (en) | Communication resource allocation method and device for satellite Internet and readable storage medium | |
Xu et al. | Resource management of GEO relays for real-time remote sensing | |
CN114884953A (en) | Internet of vehicles resource allocation method based on MEC cooperation | |
KR102439426B1 (en) | Multiple federated learning service orchestrator in edge computing and method of performing thereof | |
CN114980160A (en) | Unmanned aerial vehicle-assisted terahertz communication network joint optimization method and device | |
Lu et al. | Enhancing vehicular edge computing system through cooperative computation offloading | |
CN114827152B (en) | Low-delay cloud edge-side cooperative computing method and device for satellite-ground cooperative network | |
Yuan et al. | Actor-critic learning-based energy optimization for UAV access and backhaul networks | |
Wang et al. | Resource allocation based on Radio Intelligence Controller for Open RAN towards 6G | |
Zhang et al. | Cybertwin-driven multi-intelligent reflecting surfaces aided vehicular edge computing leveraged by deep 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 |