CN112653500B - Low-orbit satellite edge calculation-oriented task scheduling method based on ant colony algorithm - Google Patents
Low-orbit satellite edge calculation-oriented task scheduling method based on ant colony algorithm Download PDFInfo
- Publication number
- CN112653500B CN112653500B CN202011485598.XA CN202011485598A CN112653500B CN 112653500 B CN112653500 B CN 112653500B CN 202011485598 A CN202011485598 A CN 202011485598A CN 112653500 B CN112653500 B CN 112653500B
- Authority
- CN
- China
- Prior art keywords
- task
- model
- calculation
- energy consumption
- satellite
- 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
- H04B—TRANSMISSION
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/14—Relay systems
- H04B7/15—Active relay systems
- H04B7/185—Space-based or airborne stations; Stations for satellite systems
- H04B7/1851—Systems using a satellite or space-based relay
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/48—Program initiating; Program switching, e.g. by interrupt
- G06F9/4806—Task transfer initiation or dispatching
- G06F9/4843—Task transfer initiation or dispatching by program, e.g. task dispatcher, supervisor, operating system
- G06F9/4881—Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
- G06F9/4893—Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues taking into account power or heat criteria
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
- G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/46—Multiprogramming arrangements
- G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
- G06F9/5061—Partitioning or combining of resources
- G06F9/5072—Grid computing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/24—Traffic characterised by specific attributes, e.g. priority or QoS
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Astronomy & Astrophysics (AREA)
- Aviation & Aerospace Engineering (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Radio Relay Systems (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention discloses a low-orbit satellite edge computing-oriented task scheduling method based on an ant colony algorithm, which is oriented to an LEO satellite edge computing scene, establishes a system cost function of time delay and energy consumption under the constraint of computing resources (such as central processing unit frequency and a memory) of equipment, limited battery energy and different quality of service (QoS) requirements of multiple users, designs a task scheduling method based on the ant colony algorithm, optimizes the task execution sequence problem of the multiple equipment by adopting the ant colony algorithm, and optimizes the cost of local computing by scheduling clock frequency, thereby achieving the minimum total cost of a system. Simulation results show that the system cost of the algorithm is 17.5%, 14.3% and 22.2% lower than that of a random sorting algorithm, a large task first sorting algorithm and a small task first sorting algorithm respectively.
Description
Technical Field
The invention relates to the technical field of Low Earth Orbit (LEO) satellites and Mobile Edge Computing (MEC), in particular to a low earth orbit satellite-oriented edge computing task scheduling method based on an ant colony algorithm.
Background
With the commercialization of fifth Generation (5G) mobile communication networks, 6G has been required to provide more diverse services in terms of coverage and capacity. Facing to the wide coverage requirement of 6G, the satellite communication has natural advantages and can play a key role. LEO satellite networks are currently gaining high attention in academia and industry, because LEO satellites feature high capacity (1Tbps) and low latency (20-50 ms).
According to the white paper published by the European Telecommunications Standards Institute (ETSI) in 2014, MEC is defined as: "Mobile edge computing is capable of providing IT service environments and cloud computing capabilities within a Radio Access Network (RAN) in the vicinity of a mobile device. "the main research directions of MEC include: MEC service orchestration, security mechanisms, computation offload, edge caching.
For future communication demands, in desert, forest, ocean and other areas, a large number of user data processing demands of various types exist, such as data processing of human scientific research works. Meanwhile, these regional Internet of Things (IoT) terminals also need to be widely deployed, such as real-time detection and feedback of terrestrial-source marine pollutants. But these areas lack the effective coverage of terrestrial mobile communications networks, which results in that typical MEC techniques relying on terrestrial mobile communications networks cannot be applied in such scenarios. Therefore, research on satellite edge calculation has become a trend in recent years.
Disclosure of Invention
The invention aims to face LEO satellite edge computing scenes, and not only ensures reasonable distribution of limited computing resources, but also ensures that the cost of the whole system is minimum under the condition that the QoS requirement of single equipment is met by considering the aspects of MEC server resource limitation and the QoS requirement of the single equipment. The inventive content includes (1) establishing a system cost function of latency and energy consumption under the constraints of limited computing resources (central processing unit frequency and memory) of local devices, limited battery energy, and multiple devices with different QoS requirements. (2) The technical scheme for realizing the purpose of the invention by providing a calculation unloading strategy based on an ant colony algorithm, optimizing the task execution sequence problem of multiple devices by adopting the ant colony algorithm and optimizing the cost of local calculation by scheduling clock frequency to minimize the time delay and the total cost of an energy consumption system is as follows:
a low-orbit satellite edge computing-oriented task scheduling method based on an ant colony algorithm comprises the following steps:
1) establishing a communication model: according to the number K of devices (K is epsilon {1, 2.. K }), the channel bandwidth B and the transmission power P k White gaussian noise signal power σ 0 Channel gain g k,s And interference generated by other devices to the device k
Establishing an uplink transmission rate R k Model, expression is
2) Establishing a calculation model: including the required number of CPU cycles c k Uplink data size d k The calculation model is divided into a local calculation model and a satellite calculation model, and comprises the following steps:
2-1) establishing a local computation model, comprising the following steps:
2-1-1) number of CPU cycles required for utilization c k And clock frequency f k,l Establishing a local computation time delay model T k,l The expression is
2-1-2) number of CPU cycles required for utilization c k Clock frequency f k,l Establishing the calculation energy consumption model E with the effective switch capacitor kappa k,l The expression is E k,l ＝kc k (f k,l ) 2 ；
2-1-3) establishing a local computation time delay and energy consumption function psi by utilizing a local computation time delay model and a local computation energy consumption model k,l The expression is psi k,l ＝(1-β)T k,l +βE k,l Wherein β ∈ [0,1 ]]Is the weight of the time delay and energy consumption of the service;
2-2) building a satellite calculation model, including an unloading task and a satellite executing task, and specifically comprising the following steps:
2-2-1) task data size d according to k-th bit arranged in order q q(k) And a task uplink transmission rate R of the k-th bit arranged in the order q q(k) Establishing a model of the transmission delay of the satellite calculation
The expression is
2-2-2) number of CPU cycles required according to the k-th bit arranged in the order q c q(k) And satellite clock frequency f s Establishing a calculation time delay model of satellite calculation
The expression is
2-2-3) at the moment H when satellite calculation begins according to task q (k) q(k) And task q (k-1) is completed at satellite calculation completion time H q(k-1) And establishing a model for calculating the waiting time delay of the satellite
The expression is
2-2-4) establishing a transmission delay model according to the step 2-2-1)
And a device transmission power P of a k-th bit arranged in the order q q(k) Establishing a transmission energy consumption model
2-2-5) the required number of CPU cycles c for the k-th bit ordered in the order q, according to the effective switched capacitance k q(k) And the clock frequency f of the satellite s Establishment of a satellite stationModel of energy consumption kc of consumption q(k) (f s ) 2 ；
2-2-6) establishing a transmission delay model according to the step 2-2-1)
Calculating time delay model established in step 2-2-2)
Step 2-2-3) of establishing a waiting time delay model
The transmission energy consumption model established in the step 2-2-4) and the energy consumption E consumed by the satellite established in the step 2-2-5) q(k),s And establishing a total time delay and energy consumption function model psi during satellite calculation q(k),s The expression is
Wherein beta is epsilon [0,1 ∈ ]]Is a weighting factor for energy consumption and time delay;
3) establishing a problem model: the task at the device is locally calculated not to exceed the maximum time of tolerance
The energy consumption cannot exceed the available energy of the equipment
Clock frequency of local device
And the off-loading task cannot exceed the deadline
Under the constraint of (c), and to represent the offload decision o in binary k E.g. {0,1}, ensuring that the computing resources, the battery energy and the multiple users of the equipment have different QoS requirements through the constraint, and establishing a local computing time delay and energy consumption function psi according to the step 2-1-3) k,l And step 2-2-6) establishmentModel psi of the total time delay and energy consumption function in satellite calculation q(k),s Establishing a model of the minimum cost of the time delay and energy consumption system
4) The task scheduling method based on the ant colony algorithm is designed, the ant colony algorithm is adopted to optimize the task execution sequence problem of multiple devices and optimize the cost of local calculation through scheduling clock frequency, so that the total cost of a time delay and an energy consumption system is minimized, and the specific steps are as follows:
4-1) scheduling a clock frequency to optimize local computation, comprising the steps of:
4-1-1) according to the time delay and energy consumption system minimum cost module, the model for scheduling clock frequency to optimize local calculation is as follows:
4-1-2) order (1-. beta.) T k,l +βE k,l ＝Q k,l (f k,l ) Then Q is k,l (f k,l ) Is dependent only on f k,l Thus, Q is adjusted k,l (f k,l ) To f is paired k,l Derivative and equal to 0 to obtain
4-1-3) according to the constraint conditions C1 and C2 in the step 4-1-1), simplifying to obtain
And
4-1-4) combining the result of step 4-1-3) with the constraint C3 of step 4-1-1)
And
4-1-5) ensuring f k,l If the feasible region of (b) is not empty, then f is satisfied u '≤f h ', the best result of the step 4-1-1) problem is therefore:
4-2) optimizing the task execution sequence, and optimizing the task execution sequence of the multiple devices by using an ant colony algorithm, so that the model of the minimum cost of the time delay and energy consumption system in the step 3) is simplified as follows:
C2:o k ∈{0,1}
the task scheduling method based on the ant colony algorithm comprises the following steps:
4-2-1) initialization variables: before calculation, initializing the number of CPU cycles, effective switch capacitors, transmission power of equipment, data size, weight factors of energy consumption and time delay, clock frequency of a satellite, the number of ants, pheromone volatilization coefficients, maximum iteration times and related parameters of pheromone strength;
4-2-2) inputting the execution time of each task in the cloud, the transmission time, the maximum tolerance time of the calculation unloading, the system cost (except the waiting time required by the task) of the calculation unloading and the system cost of local calculation;
4-2-3) construct a solution space, as follows:
4-2-3-1) randomly arranging K tasks, randomly placing each ant on different initial tasks, and maintaining a path memory vector for storing the tasks that the ant passes through in sequence;
4-2-3-2) selecting the next task to be selected by ants by using a roulette method in each step of constructing the task queuing sequence;
4-2-3-3) selecting the next task to be selected by the ants with roulette in each step of constructing the task queuing sequence;
4-2-3-4) judging whether a task to be selected exists, if so, repeating the step 4-2-3-1), and if not, recording the optimal execution sequence of the task of the iteration;
4-2-4) record the system minimum cost for this iteration as follows:
4-2-4-1) initializing the cost of the path taken by each ant to be zero, taking out the path of each ant in sequence, and judging whether each task has waiting time;
4-2-4-2) judging whether the task of the user is calculated and unloaded according to the QoS requirement of the user;
4-2-4-3) if the QoS requirement of the user is not met, the user executes local calculation, and the time delay and the energy consumption of the local calculation are reduced by optimizing the local calculation through scheduling the clock frequency, so that the total cost of the system is reduced;
4-2-4-4) if the QoS requirement of the user is met, the user executes satellite calculation and calculates the total time delay and energy consumption calculated by the task satellite;
4-2-4-5) recording the total system cost of all the tasks of the iteration;
4-2-5) recording the task execution sequence with the minimum system cost in the current iteration times, and updating the pheromone concentration on each task execution sequence;
4-2-6) if the iteration times are smaller than the maximum iteration times, adding one to the iteration times, emptying the record table of the task execution sequence of the ants, and returning to the step 4-2-3-1), otherwise, stopping the calculation and outputting the optimal solution.
The invention provides an ant colony algorithm-based low-orbit satellite edge computing-oriented task scheduling method, which is oriented to an LEO satellite edge computing scene, establishes a time delay and energy consumption system cost function under the constraints of computing resources (such as central processing unit frequency and a memory) of equipment, battery energy limitation and different quality of service (QoS) requirements of multiple users, designs an ant colony algorithm-based task scheduling method, optimizes task execution sequence problems of multiple equipment by adopting the ant colony algorithm, and optimizes local computing cost by scheduling clock frequency, thereby achieving the minimum total cost of the system. Simulation results show that the system cost of the algorithm is 17.5%, 14.3% and 22.2% lower than that of a random ordering algorithm, a large task first ordering algorithm and a small task first ordering algorithm respectively.
Drawings
Fig. 1 is a block diagram of an MEC enhanced LEO satellite network in an embodiment;
FIG. 2 is a flowchart of a task scheduling method based on the ant colony algorithm;
FIG. 3 is a graph showing the system cost under each algorithm for the number of tasks of the apparatus used in the examples;
FIG. 4 is a graph of the system delay under each algorithm for the number of tasks of the apparatus used in the examples;
FIG. 5 is a TSM-ACA calculated system cost graph of the number of tasks of the device considering different QoS requirements of users in the embodiment.
Detailed Description
The invention will be further elucidated with reference to the drawings and examples, without however being limited thereto.
Example (b):
and setting a scene of the MEC enhanced LEO satellite network, wherein the scene comprises an LEO satellite provided with the MEC server and a plurality of devices (such as terminal devices and IoT devices), and a device K (K is in the range of {1, 2.. K }). The scenario of the MEC enhanced LEO satellite network is shown in fig. 1.
The present embodiment meets the following preconditions:
the invention provides a technical scheme based on the following preconditions:
(1) the device's tasks can be both locally computed and offloaded to satellite computing through satellite-to-ground communications.
(2) Each device has a task.
(3) The result of the downstream transmission is ignored because the size of the downstream transmitted data is much smaller than the size of the calculation input data.
A low-orbit satellite edge computing-oriented task scheduling method based on an ant colony algorithm comprises the following steps:
1) establishing a communication model: according to the number K of devices (K is equal to {1, 2.. K }), the channel bandwidth B and the transmission power P k White gaussian noise signal power σ 0 Channel gain g k,s And interference generated by other devices to the device k
Establishing an uplink transmission rate R k Model, expression is
2) Establishing a calculation model: including the required number of CPU cycles c k Uplink data size d k The calculation model is divided into a local calculation model and a satellite calculation model, and comprises the following steps:
2-1) establishing a local computation model, comprising the following steps:
2-1-1) number of CPU cycles required for utilization c k And clock frequency f k,l Establishing a local computation time delay model T k,l The expression is
2-1-2) number of CPU cycles required for utilization c k Clock frequency f k,l Establishing the calculation energy consumption model E with the effective switch capacitor kappa k,l The expression is E k,l ＝κc k (f k,l ) 2 ；
2-1-3) establishing a local computation time delay and energy consumption function psi by utilizing a local computation time delay model and a local computation energy consumption model k,l The expression is psi k,l ＝(1-β)T k,l +βE k,l Wherein beta is [0,1 ]]Is the weight of the time delay and energy consumption of the service;
2-2) building a satellite computation model, including unloading tasks and satellite executing tasks, wherein the sequence of the satellite executing tasks in the model can be represented by q ═ ((q (1)), (q (2)), (q (k))), and the specific steps are as follows:
2-2-1) task data size d according to k-th bit arranged in order q q(k) And a task uplink transmission rate R of the k-th bit arranged in the order q q(k) Establishing a model of the transmission delay of the satellite calculation
The expression is
2-2-2) number of CPU cycles required according to the k-th bit arranged in the order q c q(k) And satellite clock frequency f s Establishing a calculation time delay model of satellite calculation
The expression is
In the model, as the k-th task ranked in the order q is uploaded to the satellite, the satellite allocates all resources to the task;
2-2-3) at the moment H at which satellite calculation begins according to task q (k) q(k) And task q (k-1) at satellite calculation completion time H q(k-1) Establishing a model of satellite calculation waiting time delay
The expression is
2-2-4) establishing a transmission time delay model according to the step 2-2-1)
And a device transmission power P of a k-th bit arranged in the order q q(k) The transmission energy consumption model is established as
2-2-5) required number of CPU cycles c for the k-th bit ordered in order q according to the effective switched capacitance k q(k) And the clock frequency f of the satellite s The established energy consumption model of the satellite is kappa c q(k) (f s ) 2 ；
2-2-6) establishing a transmission time delay model according to the step 2-2-1)
Calculating time delay model established in step 2-2-2)
Step 2-2-3) established waiting time delay model
Transmission energy consumption model established in step 2-2-4)
And the energy consumption k c consumed by the satellite established in step 2-2-5) q(k) (f s ) 2 And establishing a model of the total time delay and energy consumption function in satellite calculation
Wherein beta is epsilon [0,1 ∈ ]]Is a weighting factor for energy consumption and time delay;
3) establishing a problem model: the task at the device is locally calculated not to exceed the maximum time of tolerance
The energy consumption cannot exceed the available energy of the equipment
Clock frequency of local device
And the off-loading task cannot exceed the deadline
And binary representation of the offload decision o k E {0,1}, by which constraints are ensured that the computational resources of the device (e.g., central processing unit frequency and memory), battery power, and multiuser have different QoS requirements, the local computation delay and energy consumption function ψ, established according to steps 2-1-3) k,l And the model psi of the total time delay and energy consumption function in the satellite calculation established in the step 2-2-6) q(k),s Establishing a time delay and energy consumption system minimum cost model as follows:
4) a Task scheduling method (TSM-ACA) based on the ant colony algorithm is designed, the process is shown in fig. 2, the ant colony algorithm is adopted to optimize the Task execution order problem of multiple devices and optimize the cost of local calculation by scheduling clock frequency, so that the total cost of a time delay and energy consumption system is minimized, and the specific steps are as follows:
4-1) scheduling a clock frequency to optimize local computation, comprising the steps of:
4-1-1) according to the time delay and energy consumption system minimum cost module, the model for scheduling clock frequency to optimize local calculation is as follows:
4-1-2) order (1-. beta.) T k,l +βE k,l ＝Q k,l (f k,l ) Then Q is obtained k,l (f k,l ) Is only dependent on f k,l Thus, Q will be k,l (f k,l ) To f k,l Derivative and equal to 0 to obtain
4-1-3) according to the constraint conditions C1 and C2 in the step 4-1-1), simplifying to obtain
And
4-1-4) combining the results of step 4-1-3) with the constraint C3 of step 4-1-1)
And
4-1-5) ensuring f k,l If the feasible region of (1) is not empty, then f is satisfied u '≤f h ', therefore the best results for the step 4-1-1) problem are:
4-2) optimizing the task execution sequence, and optimizing the task execution sequence of the multiple devices by adopting an ant colony algorithm, so that the model of the time delay and the minimum cost of the energy consumption system in the step 3) is simplified as follows:
C2:o k ∈{0,1}
the task scheduling method based on the ant colony algorithm comprises the following steps:
4-2-1) initialization variables: before calculation, initializing the number of CPU cycles, effective switch capacitors, transmission power of equipment, data size, weight factors of energy consumption and time delay, clock frequency of a satellite, the number of ants, pheromone volatilization coefficients, maximum iteration times and related parameters of pheromone strength;
4-2-2) inputting the execution time of each task in the cloud end, the transmission time, the maximum tolerance time of calculation unloading, the system cost of calculation unloading (except the waiting time required by the task) and the system cost of local calculation;
4-2-3) construct a solution space, as follows:
4-2-3-1) randomly arranging K tasks, randomly placing each ant on different initial tasks, and maintaining a path memory vector for storing the tasks that the ant passes through in sequence;
4-2-3-2) the ants select the next task to be selected by roulette in each step of constructing the task queuing sequence;
4-2-3-3) selecting the next task to be selected by the ant by using a roulette method in each step of constructing the task queuing sequence;
4-2-3-4) judging whether a task to be selected exists, if so, repeating the step 4.2.3, and if not, recording the optimal execution sequence of the task of the iteration.
4-2-4) record the system minimum cost for this iteration as follows:
4-2-4-1) initializing the cost of the path taken by each ant to be zero, taking out the path of each ant in sequence, and judging whether the waiting time exists in each task;
4-2-4-2) judging whether the task of the user is calculated and unloaded according to the QoS requirement of the user;
4-2-4-3) if the QoS requirement of the user is not met, the user executes local calculation, and the time delay and the energy consumption of the local calculation are reduced by optimizing the local calculation through scheduling the clock frequency, so that the total cost of the system is reduced;
4-2-4-4) if the QoS requirement of the user is met, the user executes satellite calculation and calculates the total time delay and energy consumption calculated by the task satellite;
4-2-4-5) recording the total system cost of all the tasks of the iteration;
4-2-5) recording the task execution sequence with the minimum system cost in the current iteration times, and meanwhile, updating the pheromone concentration on each task execution sequence;
4-2-6) if the iteration times is less than the maximum iteration times, adding one to the iteration times, emptying the record table of the ant passing through the task execution sequence, and returning to the step 4.2.3, otherwise, terminating the calculation and outputting the optimal solution.
The method is adopted for simulation, the performance of the TSM-ACA is verified, and the TSM-ACA is compared with the following unloading scheme: random ordering (random queue), small task first ordering (SDFQ), large task first ordering (LTFQ).
The simulation parameters are set as follows: uplink transmission rate R k 10 MBps; number of CPU cycles c k ＝[0.1×10 9 ,1×10 9 ]cycles; time tolerated by local computing tasks
Effective switched capacitor k 10 -26 (ii) a Device transmission power P q(k) 20 mW; uplink data size d k ＝[2,6]MB; clock frequency f of satellite s ＝[6,8]GC/s(GC＝10 9 cycles)。
As shown in FIG. 3, the system cost of the algorithm of the present invention is 17.5%, 14.3%, and 22.2% lower than the random ordering algorithm, the big task first ordering algorithm, and the little task first ordering algorithm, respectively. As the number of tasks of the device increases, the system cost of the three algorithms increases because competition between tasks becomes more and more intense, and therefore the cost of the system is linearly related to the number of devices; meanwhile, the value of the cost of the small task first sorting algorithm is very close to that of the random sorting algorithm system, which shows that the number of the equipment tasks unloaded by the two algorithms is almost close.
As shown in fig. 4, four algorithms comparing the task numbers of different devices are compared, and as the task number of a device increases, the system delay cost of the four algorithms also increases; obviously, under the TSM-ACA algorithm, the delay cost of the system is always the lowest, and the delay cost of the random ordering algorithm, the large task first ordering algorithm and the small task first ordering algorithm is very close to that of the system; the delay cost of the algorithm system is respectively 17.5%, 16.5% and 22.7% lower than that of a random sorting algorithm, a large task first sorting algorithm and a small task first sorting algorithm; at the same time, it can be seen from fig. 4 and 5 that the major cost in the overall system is due in large part to energy consumption.
As shown in fig. 5, the simulation is performed under the TSM-ACA algorithm, and the three curves in the diagram use the costs of the system with different device task quantities respectively under the condition of calculating the unloading deadlines of 2.5s, 3s and 3.5s, namely representing different user QoS requirements. It can be seen from the figure that the overall system cost increases as the number of device tasks increases for different users 'QoS requirements, since the number of tasks representing a device is more counted at the satellite when the user's QoS requirements are lower.
Claims (2)
1. A low-orbit satellite edge computing-oriented task scheduling method based on an ant colony algorithm is characterized by comprising the following steps:
1) establishing a communication model: according to the number K of devices (K is epsilon {1, 2.. K }), the channel bandwidth B and the transmission power P k White gaussian noise signalPower of σ 0 Channel gain g k,s And interference generated by other devices to device k
Establishing an uplink transmission rate R k Model, expression is
2) Establishing a calculation model: including the required number of CPU cycles c k Uplink data size d k The calculation model is divided into a local calculation model and a satellite calculation model, and comprises the following steps:
2-1) establishing a local computation model, comprising the following steps:
2-1-1) number of CPU cycles required for utilization c k And clock frequency f k,l Establishing a local computation time delay model T k,l Of the formula
2-1-2) number of CPU cycles required for utilization c k Clock frequency f k,l Establishing the calculation energy consumption model E with the effective switch capacitor kappa k,l The expression is E k,l ＝κc k (f k,l ) 2 ；
2-1-3) establishing a local computation time delay and energy consumption function psi by utilizing a local computation time delay model and a local computation energy consumption model k,l The expression is psi k,l ＝(1-β)T k,l +βE k,l Wherein beta is [0,1 ]]The weight of the time delay and the energy consumption of the service;
2-2) building a satellite calculation model, including an unloading task and a satellite execution task, and specifically comprising the following steps:
2-2-1) size d of task data according to k-th bit arranged in order q q(k) And a task uplink transmission rate R of the k-th bit arranged in the order q q(k) Establishing a model of the transmission delay of the satellite calculation
The expression is
2-2-2) number of CPU cycles required according to the k-th bit arranged in the order q c q(k) And satellite clock frequency f s Establishing a calculation time delay model of satellite calculation
The expression is
2-2-3) at the moment H at which satellite calculation begins according to task q (k) q(k) And task q (k-1) is completed at satellite calculation completion time H q(k-1 And establishing a model for calculating the waiting time delay of the satellite
The expression is
2-2-4) establishing a transmission time delay model according to the step 2-2-1)
And a device transmission power P of a k-th bit arranged in the order q q(k) Establishing a transmission energy consumption model
2-2-5) the required number of CPU cycles c for the k-th bit ordered in the order q according to the effective switched capacitance k q(k) And the clock frequency f of the satellite s And establishing an energy consumption model k c consumed by the satellite q(k) (f s ) 2 ；
2-2-6) establishing a transmission delay model according to the step 2-2-1)
Calculating time delay model established in step 2-2-2)
Step 2-2-3) established waiting time delay model
The transmission energy consumption model established in the step 2-2-4) and the energy consumption E consumed by the satellite established in the step 2-2-5) q(k),s And establishing a total time delay and energy consumption function model psi during satellite calculation q(k),s The expression is
Wherein beta is epsilon [0,1 ∈ ]]Is a weighting factor for energy consumption and time delay;
3) establishing a problem model: the task at the device is locally calculated not to exceed the maximum time of tolerance
The energy consumption cannot exceed the available energy of the equipment
Clock frequency of local device
And the off-loading task cannot exceed the deadline
And binary representation of the offload decision o k E {0,1}, ensuring that the computing resources, the battery energy and the multi-user of the equipment have different QoS requirements through the constraint, and establishing a local computing time delay and energy consumption function psi according to the step 2-1-3) k,l And the model psi of the total time delay and energy consumption function in the satellite calculation established in the step 2-2-6) q(k),s Establishing a model of the minimum cost of the time delay and energy consumption system
4) The task scheduling method based on the ant colony algorithm is designed, the ant colony algorithm is adopted to optimize the task execution sequence problem of the multiple devices, and the cost of local calculation is optimized by scheduling the clock frequency, so that the time delay and the total cost of an energy consumption system are minimized.
2. The method for scheduling the low-orbit satellite edge calculation task based on the ant colony algorithm as claimed in claim 1, wherein in the step 4), the method for scheduling the task based on the ant colony algorithm comprises the following specific steps:
4-1) scheduling clock frequency optimization local computation, comprising the steps of:
4-1-1) according to the time delay and energy consumption system minimum cost module, the model for scheduling clock frequency to optimize local calculation is as follows:
4-1-2) order
Then Q is k,l (f k,l ) Is only dependent onf k,l Thus, Q will be k,l (f k,l ) To f is paired k,l Derivative and equal to 0 to obtain
4-1-3) according to the constraint conditions C1 and C2 in the step 4-1-1), simplifying to obtain
And
4-1-4) combining the results of step 4-1-3) with the constraint C3 of step 4-1-1)
And
4-1-5) ensuring f k,l If the feasible region of (b) is not empty, then f is satisfied u '≤f h ', the best result of the step 4-1-1) problem is therefore:
4-2) optimizing the task execution sequence, and optimizing the task execution sequence of the multiple devices by using an ant colony algorithm, so that the model of the minimum cost of the time delay and energy consumption system in the step 3) is simplified as follows:
C2:o k ∈{0,1}
the task scheduling method based on the ant colony algorithm comprises the following steps:
4-2-1) initialization variables: before calculation, initializing the number of CPU cycles, effective switch capacitors, transmission power of equipment, data size, weight factors of energy consumption and time delay, clock frequency of a satellite, the number of ants, pheromone volatilization coefficients, maximum iteration times and related parameters of pheromone strength;
4-2-2) inputting the execution time of each task at the cloud end, the transmission time, the maximum tolerance time of calculation unloading, the system cost of calculation unloading and the system cost of local calculation;
4-2-3) construct a solution space as follows:
4-2-3-1) randomly arranging K tasks, randomly placing each ant on different initial tasks, and maintaining a path memory vector for storing the tasks that the ant passes through in sequence;
4-2-3-2) selecting the next task to be selected by the ant by using a roulette method in each step of constructing the task queuing sequence;
4-2-3-3) the ants select the next task to be selected by roulette in each step of constructing the task queuing sequence;
4-2-3-4) judging whether a task to be selected exists, if so, repeating the step 4-2-3-1), and if not, recording the optimal execution sequence of the task of the iteration;
4-2-4) record the system minimum cost for this iteration as follows:
4-2-4-1) initializing the cost of the path taken by each ant to be zero, taking out the path of each ant in sequence, and judging whether the waiting time exists in each task;
4-2-4-2) judging whether the task of the user is calculated and unloaded according to the QoS requirement of the user;
4-2-4-3) if the QoS requirement of the user is not met, the user executes local calculation, and the local calculation is optimized by scheduling clock frequency to reduce the time delay and energy consumption of the local calculation and reduce the total cost of the system;
4-2-4-4) if the QoS requirement of the user is met, the user executes satellite calculation and calculates the total time delay and energy consumption calculated by the task satellite;
4-2-4-5) recording the total system cost of all the tasks of the iteration;
4-2-5) recording the task execution sequence with the minimum system cost in the current iteration times, and updating the pheromone concentration on each task execution sequence;
4-2-6) if the iteration times are less than the maximum iteration times, adding one to the iteration times, emptying the record table of the ant passing through the task execution sequence, and returning to the step 4-2-3-1), otherwise, terminating the calculation and outputting the optimal solution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011485598.XA CN112653500B (en) | 2020-12-16 | 2020-12-16 | Low-orbit satellite edge calculation-oriented task scheduling method based on ant colony algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011485598.XA CN112653500B (en) | 2020-12-16 | 2020-12-16 | Low-orbit satellite edge calculation-oriented task scheduling method based on ant colony algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112653500A CN112653500A (en) | 2021-04-13 |
| CN112653500B true CN112653500B (en) | 2022-07-26 |
Family
ID=75354214
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011485598.XA Active CN112653500B (en) | 2020-12-16 | 2020-12-16 | Low-orbit satellite edge calculation-oriented task scheduling method based on ant colony algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112653500B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113794494B (en) * | 2021-07-30 | 2023-03-28 | 湖北微源卓越科技有限公司 | Edge computing system and computing unloading optimization method for low-orbit satellite network |
| CN114124195B (en) * | 2021-11-19 | 2024-10-11 | 桂林电子科技大学 | Multi-star MEC calculation unloading strategy for LEO satellite system |
| CN114268357B (en) * | 2021-11-28 | 2023-04-25 | 西安电子科技大学 | Method, system, equipment and application for unloading computing tasks based on low-orbit satellite edges |
| CN114048040B (en) * | 2021-11-29 | 2022-05-13 | 中南大学 | Task scheduling method based on time delay relation between memory and image classification model |
| CN115250142B (en) * | 2021-12-31 | 2023-12-05 | 中国科学院上海微系统与信息技术研究所 | Star-earth fusion network multi-node computing resource allocation method based on deep reinforcement learning |
| CN114385272B (en) * | 2022-03-24 | 2022-07-05 | 山东省计算中心（国家超级计算济南中心） | Ocean task oriented online adaptive computing unloading method and system |
| CN114928394B (en) * | 2022-04-06 | 2024-08-02 | 中国科学院上海微系统与信息技术研究所 | Low-orbit satellite edge computing resource allocation method with optimized energy consumption |
| CN114513247B (en) * | 2022-04-18 | 2022-07-22 | 北京开运联合信息技术集团股份有限公司 | Method for distributing multi-satellite control data information |
| CN117200873B (en) * | 2023-11-07 | 2024-05-31 | 南京邮电大学 | Calculation unloading method considering satellite mobility in satellite edge calculation network |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103345504A (en) * | 2013-07-03 | 2013-10-09 | 邢立宁 | Operator construction method of single-star scheduling |
| CN104361234A (en) * | 2014-11-15 | 2015-02-18 | 北京理工大学 | Method for optimizing multi-star multitask observation dispatching under complicated constraint condition |
| CN106648852A (en) * | 2016-11-18 | 2017-05-10 | 合肥工业大学 | Multi-satellite mission scheduling method and device based on double ant colonies |
| CN108776612A (en) * | 2018-04-11 | 2018-11-09 | 深圳大学 | A kind of cloud computing method for allocating tasks, device, equipment and storage medium |
| CN109299490A (en) * | 2018-05-07 | 2019-02-01 | 湖南航升卫星科技有限公司 | Mission planning method is stared over the ground based on band introduce taboo list ant group algorithm video satellite |
| CN111913788A (en) * | 2020-06-19 | 2020-11-10 | 合肥工业大学 | Task scheduling method and system for imaging satellite |
-
2020
- 2020-12-16 CN CN202011485598.XA patent/CN112653500B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103345504A (en) * | 2013-07-03 | 2013-10-09 | 邢立宁 | Operator construction method of single-star scheduling |
| CN104361234A (en) * | 2014-11-15 | 2015-02-18 | 北京理工大学 | Method for optimizing multi-star multitask observation dispatching under complicated constraint condition |
| CN106648852A (en) * | 2016-11-18 | 2017-05-10 | 合肥工业大学 | Multi-satellite mission scheduling method and device based on double ant colonies |
| CN108776612A (en) * | 2018-04-11 | 2018-11-09 | 深圳大学 | A kind of cloud computing method for allocating tasks, device, equipment and storage medium |
| WO2019196127A1 (en) * | 2018-04-11 | 2019-10-17 | 深圳大学 | Cloud computing task allocation method and apparatus, device, and storage medium |
| CN109299490A (en) * | 2018-05-07 | 2019-02-01 | 湖南航升卫星科技有限公司 | Mission planning method is stared over the ground based on band introduce taboo list ant group algorithm video satellite |
| CN111913788A (en) * | 2020-06-19 | 2020-11-10 | 合肥工业大学 | Task scheduling method and system for imaging satellite |
Non-Patent Citations (1)
| Title |
|---|
| 成像卫星调度问题研究;余堃;《中国优秀硕士学位论文全文数据库（电子期刊）》;20190115;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112653500A (en) | 2021-04-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112653500B (en) | Low-orbit satellite edge calculation-oriented task scheduling method based on ant colony algorithm | |
| Bi et al. | Joint optimization of service caching placement and computation offloading in mobile edge computing systems | |
| CN112492626B (en) | Method for unloading computing task of mobile user | |
| CN110928654B (en) | Distributed online task unloading scheduling method in edge computing system | |
| CN110941667A (en) | Method and system for calculating and unloading in mobile edge calculation network | |
| CN110096362B (en) | Multitask unloading method based on edge server cooperation | |
| CN110493360A (en) | The mobile edge calculations discharging method of system energy consumption is reduced under multiserver | |
| CN113286317B (en) | Task scheduling method based on wireless energy supply edge network | |
| CN114697333B (en) | Edge computing method for energy queue equalization | |
| CN113590279B (en) | Task scheduling and resource allocation method for multi-core edge computing server | |
| CN111711962A (en) | Cooperative scheduling method for subtasks of mobile edge computing system | |
| CN113573363B (en) | MEC calculation unloading and resource allocation method based on deep reinforcement learning | |
| CN115396953B (en) | Calculation unloading method based on improved particle swarm algorithm in mobile edge calculation | |
| CN113946423A (en) | Multi-task edge computing scheduling optimization method based on graph attention network | |
| CN112416603A (en) | Combined optimization system and method based on fog calculation | |
| CN115297013A (en) | Task unloading and service cache joint optimization method based on edge cooperation | |
| CN114025359B (en) | Resource allocation and calculation unloading method, system, equipment and medium based on deep reinforcement learning | |
| CN115119234A (en) | Method for optimizing task processing of wireless equipment in wireless energy supply edge computing network | |
| CN116668439A (en) | Edge computing task unloading method based on deep reinforcement learning | |
| CN118250750A (en) | Satellite edge computing task unloading and resource allocation method based on deep reinforcement learning | |
| CN113507712A (en) | Resource allocation and calculation task unloading method based on alternative direction multiplier | |
| CN116887344A (en) | Cloud edge end collaborative system optimization method and device based on improved binary multi-universe algorithm | |
| CN115022893B (en) | Resource allocation method for minimizing total computation time in multi-task edge computing system | |
| CN116761218A (en) | Combined optimization method, system and storage medium in cloud edge end cooperative system | |
| CN115134364B (en) | Energy-saving computing and unloading system and method based on O-RAN (O-radio Access network) Internet of things system |
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 |