CN113472427B - Satellite network queue management method based on flow prediction - Google Patents

Satellite network queue management method based on flow prediction Download PDF

Info

Publication number
CN113472427B
CN113472427B CN202110752131.5A CN202110752131A CN113472427B CN 113472427 B CN113472427 B CN 113472427B CN 202110752131 A CN202110752131 A CN 202110752131A CN 113472427 B CN113472427 B CN 113472427B
Authority
CN
China
Prior art keywords
model
flow data
exponential smoothing
algorithm
dynamic
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
Application number
CN202110752131.5A
Other languages
Chinese (zh)
Other versions
CN113472427A (en
Inventor
别玉霞
李芷含
胡智
王宇鹏
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenyang Aerospace University
Original Assignee
Shenyang Aerospace University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shenyang Aerospace University filed Critical Shenyang Aerospace University
Priority to CN202110752131.5A priority Critical patent/CN113472427B/en
Publication of CN113472427A publication Critical patent/CN113472427A/en
Application granted granted Critical
Publication of CN113472427B publication Critical patent/CN113472427B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • H04B7/18519—Operations control, administration or maintenance
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/25—Design optimisation, verification or simulation using particle-based methods
    • 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
    • H04L41/14—Network analysis or design
    • H04L41/142—Network analysis or design using statistical or mathematical methods
    • 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
    • H04L41/14—Network analysis or design
    • H04L41/147—Network analysis or design for predicting network behaviour
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Analysis (AREA)
  • Probability & Statistics with Applications (AREA)
  • Pure & Applied Mathematics (AREA)
  • Computer Hardware Design (AREA)
  • Evolutionary Computation (AREA)
  • Geometry (AREA)
  • General Engineering & Computer Science (AREA)
  • Algebra (AREA)
  • Astronomy & Astrophysics (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

The invention provides a satellite network queue management method based on flow prediction, which constructs a dynamic cubic exponential smoothing model, takes flow data with self-similarity as the input of the dynamic cubic exponential smoothing model, optimizes a smoothing coefficient in the dynamic cubic exponential smoothing model by using a differential evolution algorithm, takes the dynamic cubic exponential smoothing model corresponding to the optimal smoothing coefficient as a prediction model, performs smooth nonlinear processing on a packet discarding probability function of an ARED algorithm by using a cubic curve function, outputs the packet discarding probability for controlling the packet loss rate, throughput and average queue length in a transmission channel by using the ARED algorithm, improves the accuracy of flow prediction, provides sufficient time for network congestion control, improves the queue length by using the nonlinear packet discarding probability function, and solves the parameter sensitivity problem of RED while effectively controlling the queue length, the channel transmission is more stable.

Description

Satellite network queue management method based on flow prediction
Technical Field
The invention belongs to the technical field of satellite network communication, and particularly relates to a satellite network queue management method based on flow prediction.
Background
The time series model is the most widely used flow prediction model and is divided into a moving average model, a stationary time series model and an exponential smoothing model. Holt in 1958, the principle is that the exponential smoothing value of any stage is the weighted average of the actual observed value of the stage and the exponential smoothing value of the previous stage, and the influence of short-term random fluctuation on the sequence is weakened by using a smoothing technology to smooth the sequence, so that the time series smoothing value is obtained and used as a prediction parameter in a short term in the future. The algorithm is the most widely used method in production prediction and is also used for predicting the medium-short term economic development trend.
The exponential smoothing model has different smoothing times, wherein a cubic exponential smoothing (Holt-Winter) prediction model mainly corrects nonlinear trends in a time sequence and can adapt to the nonlinear, self-similarity and long correlation change trend of satellite network traffic. The Holt-Winter algorithm comprises three smooth equations and a prediction equation, the observed values in each period are weighted and averaged according to the time sequence, the result is used as a flow prediction value, and the characteristic that the influence of historical data on future values is reduced along with the time is shown. The smoothing coefficient alpha of the traditional cubic exponential smoothing model is a fixed value, the model is mainly applied to a stable data model, and in medium-term and long-term data prediction, when data change is large, the data change cannot be adjusted in time, so that prediction errors become large. Therefore, some scholars improve the traditional cubic exponential smoothing method and provide a dynamic cubic exponential smoothing prediction model. The dynamic cubic exponential smoothing method is to provide a dynamic smoothing coefficient on the basis of the traditional algorithm and update parameters in an iterative mode. The algorithm can reduce the prediction error in the medium-long term prediction of complex data, and has better stability.
Typical queue management models include Random Early Detection (RED), Adaptive Random Early Detection (arm), and the like. In the RED queue management method, the sensitivity of parameter setting is high, the packet loss probability is a linear function, the queue oscillation is easily caused by the too high packet loss rate, and the channel utilization rate is too low; the ARED queue management method realizes automatic adjustment of parameters, but the packet dropping probability function in the ARED is still linear, and the queue oscillation is easily caused by the excessively fast linear increasing rate.
Disclosure of Invention
Aiming at the defects of the prior art, the invention introduces the cubic curve function to improve the ARED algorithm, can effectively reduce the packet loss rate, improve the average queue length and improve the channel utilization rate. Therefore, the invention provides a satellite network queue management method based on flow prediction, which is suitable for satellite network service flow with self-similarity characteristics, and comprises the following steps:
step 1: carrying out self-similarity judgment on flow data in a satellite network within a period of time; the concrete expression is as follows: estimating the Hurst index H of the flow data by adopting a re-standard polar difference R/S analysis method, and if H belongs to (0.5,1), indicating that the flow data has self-similarity;
and 2, step: taking the flow data with self-similarity as the input of a dynamic cubic exponential smoothing model, and combining the dynamic cubic exponential smoothing model and an ARED algorithm to obtain the packet discarding probability;
and step 3: and controlling the packet loss rate, the throughput and the average queue length in the transmission channel according to the obtained packet loss probability.
The step 2 comprises the following steps:
step 2.1: constructing a dynamic cubic exponential smoothing model, taking flow data with self-similarity as input of the dynamic cubic exponential smoothing model, and optimizing a smoothing coefficient in the dynamic cubic exponential smoothing model by using a differential evolution algorithm to obtain an optimal solution of the smoothing coefficient;
step 2.2: taking a dynamic cubic exponential smoothing model corresponding to the optimal solution of the smoothing coefficient as an optimal prediction model, and taking a predicted value of flow data output by the optimal prediction model as the input of an ARED algorithm;
step 2.3: and performing smooth nonlinear processing on the packet discarding probability function of the ARED algorithm by using a cubic curve function, and outputting the packet discarding probability through the ARED algorithm.
The step 2.1 comprises:
step 2.1.1: initializing an initial population in a differential evolution algorithm;
step 2.1.2: carrying out crossing, variation and selection on the initial population by using a differential evolution algorithm, and assigning the value corresponding to each generation of output individuals to a smoothing coefficient in a dynamic cubic exponential smoothing model after the maximum iteration times are reached;
step 2.1.3: taking the flow data with self-similarity as the input of a dynamic cubic exponential smoothing model, predicting the flow data by using the dynamic cubic exponential smoothing model, and outputting the predicted value of the flow data;
step 2.1.4: the sample mean square error f (Δ) is calculated using equation (1):
Figure BDA0003140879240000021
in the formula, xiIndicating the incoming flow data, SiA predicted value of the output time queue is represented, and I represents the total number of flow data;
step 2.1.5: changing the value range of the initial population, and executing the step 2.1.2-step 2.1.4 again;
step 2.1.6: and (4) repeating the step 2.1.5, executing N times of calculation to obtain N sample mean square errors, and taking the individual value when the sample mean square error is minimum as the optimal value of the smoothing coefficient.
In the step 2.3, the packet dropping probability function of the aired algorithm is subjected to smooth nonlinear processing by using a cubic curve function, and the method includes the following steps:
step 2.3.1: the average queue length Qavg is calculated using equation (2):
Qavg＝(1-ωq)Qavg+ωq·q (2)
in the formula, omegaqRepresents a weight value, q represents an initial queue length;
step 2.3.2: calculating the packet discard probability P using equation (3)b：
Figure BDA0003140879240000031
Where maxp represents the maximum packet drop probability, minthLower limit value, max, indicating a set queue thresholdthRepresents an upper limit value of the set queue threshold.
The beneficial effects of the invention are:
the invention provides a satellite network queue management method based on flow prediction, aiming at the self-similarity characteristic of satellite network flow, the invention completes flow prediction based on a dynamic cubic exponential smoothing model, and optimizes the smoothing coefficient of the model through a Differential Evolution (DE) algorithm, thereby improving the prediction precision; further, based on an ARED algorithm, a cubic function is utilized to carry out nonlinear processing on the packet discarding probability function; the method improves the accuracy of flow prediction, provides sufficient time for network congestion control, improves the ARED algorithm by utilizing the nonlinear packet dropping probability function, can effectively control the queue length, solves the parameter sensitivity problem of RED, ensures that channel transmission is more stable, and greatly improves the accuracy, the optimal performance and the like.
Drawings
FIG. 1 is a flow chart of a method for managing a satellite network queue based on traffic prediction according to the present invention;
FIG. 2 is a schematic diagram of a satellite network queue management method based on traffic prediction according to the present invention;
FIG. 3 is a schematic diagram of a differential evolution model in the present invention;
FIG. 4 is a comparison graph of the predicted results between the optimized dynamic cubic exponential smoothing model and the unoptimized dynamic cubic exponential smoothing model, ARMA model in the present invention;
FIG. 5 is a graph comparing the average queue length of the method of the present invention with an ARED model, and an ARED model not improved based on traffic prediction;
FIG. 6 is a graph comparing packet loss ratios of the method of the present invention and an ARED model based on traffic prediction for an unmodified ARED model;
FIG. 7 is a graph of a cubic function of the application of the improved ARED model of the present invention;
FIG. 8 is a schematic diagram of the improved packet dropping probability of the ARED model in the present invention.
Detailed Description
The invention is further described with reference to the following figures and specific examples.
As shown in fig. 1 to 2, a method for managing a satellite network queue based on traffic prediction includes:
step 1: carrying out self-similarity judgment on flow data in a satellite network within a period of time; the concrete expression is as follows: estimating the Hurst index H of the flow data by adopting a re-standard polar difference R/S analysis method, and if H is within the range of 0.5 and 1, indicating that the flow data has self-similarity;
the long correlation of the flow is expressed by a Hurst parameter value H, if the value range of the Hurst index is (0.5,1), the sequence has long-term memory, the future increment is correlated with the past increment, and the possibility of continuously keeping the existing trend is strong. And (3) estimating the Hurst parameter of the original flow data of the satellite network by using a re-standard polar difference R/S analysis method to obtain H which is 0.835, and knowing that the original data stream (namely the flow data) has higher long correlation, namely self-similarity. From the above, it can be seen that the flow data exhibiting self-similarity has predictability for predicting the future trend of the flow data.
And 2, step: taking the flow data with self-similarity as the input of a dynamic cubic exponential smoothing model, and combining the dynamic cubic exponential smoothing model and an ARED algorithm to obtain the packet discarding probability; the method comprises the following steps:
step 2.1: constructing a dynamic cubic exponential smoothing model, taking flow data with self-similarity as input of the dynamic cubic exponential smoothing model, and optimizing a smoothing coefficient in the dynamic cubic exponential smoothing model by using a differential evolution algorithm to obtain an optimal solution of the smoothing coefficient; the method comprises the following steps:
step 2.1.1: initializing an initial population in a differential evolution algorithm;
initial population
Figure BDA0003140879240000041
Randomly generating;
Figure BDA0003140879240000042
wherein x isi(0) Denotes the ith individual, x, of the 0 th generation in the populationj,i(0) The j-th "gene" of the i-th "individual" of the 0-th generation,
Figure BDA0003140879240000043
and
Figure BDA0003140879240000044
respectively the lower limit and the upper limit of the population individuals; NP denotes the population size, D denotes the dimension of the spatial solution, and rand (0,1) denotes random numbers uniformly distributed in the (0,1) interval.
Step 2.1.2: as shown in fig. 3, the initial population is crossed, varied and selected by using a differential evolution algorithm, and when the maximum iteration number is reached, the output values corresponding to each generation of individuals are respectively assigned to the smoothing coefficients in the dynamic cubic exponential smoothing model;
marking the smoothing coefficients as alpham,nM is predicted flow data, n is the time point of the data prediction, and the dynamic cubic exponential smoothing equation is as follows:
Figure BDA0003140879240000045
in the formula stWhich is representative of the raw flow data,
Figure BDA0003140879240000046
representing a first value of flow data at time t.
The smoothing coefficient and the parameter of each time point in the dynamic cubic exponential smoothing model are dynamically changed, and each period needs iterative computation, namely
Figure BDA0003140879240000055
Figure BDA0003140879240000052
Figure BDA0003140879240000053
In the formula, atAs a horizontal parameter, btAs a trend parameter, ctAs a seasonal parameter, yt+TFor the predicted value of the future T phase, the prediction equation is as follows:
yt+T＝at+btT+ctT2,T＝1,2,3,…
step 2.1.3: taking the flow data with self-similarity as the input of a dynamic cubic exponential smoothing model, predicting the flow data by using the dynamic cubic exponential smoothing model, and outputting a predicted value of the flow data as the initial queue length;
step 2.1.4: the sample mean square error f (Δ) is calculated using equation (1):
Figure BDA0003140879240000054
in the formula, xiIndicating the incoming flow data, SiA predicted value representing the output time queue, I represents the total number of flow data;
step 2.1.5: changing the value range of the initial population, and executing the step 2.1.2 to the step 2.1.4 again;
step 2.1.6: repeating the step 2.1.5, performing N times of calculation to obtain N sample mean square errors, and taking an individual value when the sample mean square errors are minimum as an optimal value of a smoothing coefficient;
step 2.2: taking a dynamic cubic exponential smoothing model corresponding to the optimal solution of the smoothing coefficient as an optimal prediction model, and taking a predicted value of flow data output by the optimal prediction model as the input of an ARED algorithm;
step 2.3: carrying out smooth nonlinear processing on a packet discarding probability function of the ARED algorithm by utilizing a cubic curve function, and outputting packet discarding probability through the ARED algorithm;
and step 3: controlling the packet loss rate, throughput, and average queue length in the transmission channel according to the obtained packet loss probability, as shown in fig. 5 and 6; a comparison graph of the prediction results of the optimized dynamic cubic exponential smoothing Model, the optimized cubic exponential smoothing Model, and the Auto Regression Moving Average (ARMA) Model in the stationary time series Model is shown in fig. 4.
In the step 2.3, the packet dropping probability function of the aired algorithm is subjected to smooth nonlinear processing by using a cubic curve function, and the method includes the following steps:
step 2.3.1: the average queue length Qavg is calculated using equation (2):
Qavg＝(1-ωq)Qavg+ωq·q (2)
in the formula, ωqRepresenting weight values, q representing initial queue length;
step 2.3.2: calculating the packet drop probability P using equation (3)b：
Figure BDA0003140879240000061
Where maxp represents the maximum packet drop probability, minthLower limit value, max, indicating a set queue thresholdthAn upper limit value representing a set queue threshold;
wherein in formula (3)th≤Qavg≤maxthThe determination process of the corresponding cubic non-linear function is as follows:
first, a general expression for the cubic nonlinear function is defined as: y (x) a0+a1x+a2x2+a3x3Wherein a is0、a1、a2、a3Is the undetermined coefficient;
constraint conditions y (0) 0, y (X) 1, y '(0) 0, and y' (X) 0 are set for the start time 0 and the end time T, respectively, and a is obtained0＝0,a1＝0,
Figure BDA0003140879240000062
Therefore, when the probability when all packets in the communication channel are dropped is set to 1, and when the probability when no packet loss occurs is 0, the coordinates of the start point of the cubic curve function are set to (0,0), and the end point is set to (X, 1), X represents the upper limit value of the queue threshold:
Figure BDA0003140879240000063
translating equation (4) to the right by a units, where a represents the lower limit of the queue threshold, yields:
Figure BDA0003140879240000064
taking the independent variable x in the formula (5) as the average queue length, and making a be minth，minthFor the lower threshold in the dynamic cubic exponential smoothing model, let X + a be maxth，maxthFor the upper threshold limit in the dynamic cubic exponential smoothing model, equation (5) is rewritten as:
Figure BDA0003140879240000071
utilizing formula (6) to discard min in probability function of packet in AREDth≤Qavg≤maxthAnd partially carrying out nonlinear improvement, as shown in fig. 7, obtaining a cubic function improved ARED packet discarding probability function, slowly discarding the data which just enters the ARED packet, and quickly discarding the data which gradually increases in the later period, so as to ensure the utilization rate of the channel.
According to an ARED algorithm, setting the target queue length target, wherein the value range of the target is [ min ]th+0.4(maxth-minth)，minth+0.6(maxth-minth)]And carrying out self-adaptive adjustment on maxp by utilizing the relation between target and Qavg. If Qavg is at the minimum threshold minthNear and Qavg<target, which indicates that the congestion adjustment is too positive, and reduces the value of maxp; if Qavg is at the maximum threshold maxthNear and Qavg>target, indicating that congestion adjustments are too conservative, increasing the value of maxp. Let maxp + and maxp-represent the maximum drop probability obtained after aggressive increase and conservative decrease, respectively, as follows:
Figure BDA0003140879240000072
where, maxp is the maximum value of the packet dropping probability, α is an increasing factor, α ═ min (0.01, maxp/4), and β is a decreasing factor, usually 0.9. Different discarding probability curves can be obtained when maxp takes different values.
Fig. 8 shows the packet dropping probability P when maxp is 0.5bCurve as a function of Qavg. When Qavg<minthWhen P is presentb0. When Qavg>minthWhen is, P b1. When minth≤Qavg≤maxthWhen is, PbChanges in a nonlinear trend, if the number of data packets in the queue is less, the Qavg is smaller, and the packet dropping probability P is lowerbThe growth is relatively slow; if the number of packets in the queue is increasing and Qavg is larger, P isbThe growth becomes fast. The invention effectively solves the problem of over-fast increase of packet loss rate by combining the ARED model with the nonlinear packet loss probability function, is not easy to cause queue oscillation, ensures more stable flow transmission and improved channel utilization rate, and simultaneously solves the parameter sensitivity of RED (reverse enhanced discovery) so as to well control network congestion.

Claims (3)

1. A satellite network queue management method based on flow prediction is characterized by comprising the following steps:
step 1: carrying out self-similarity judgment on flow data in a satellite network within a period of time; the concrete expression is as follows: estimating the Hurst index H of the flow data by adopting a re-standard polar difference R/S analysis method, and if H is within the range of 0.5 and 1, indicating that the flow data has self-similarity;
step 2: taking the flow data with self-similarity as the input of a dynamic cubic exponential smoothing model, and combining the dynamic cubic exponential smoothing model and an ARED algorithm to obtain the packet discarding probability;
and step 3: controlling the packet loss rate, the throughput and the average queue length in a transmission channel according to the obtained packet discarding probability;
the step 2 comprises the following steps:
step 2.1: constructing a dynamic cubic exponential smoothing model, taking flow data with self-similarity as input of the dynamic cubic exponential smoothing model, and optimizing a smoothing coefficient in the dynamic cubic exponential smoothing model by using a differential evolution algorithm to obtain an optimal solution of the smoothing coefficient;
step 2.2: taking a dynamic cubic exponential smoothing model corresponding to the optimal solution of the smoothing coefficient as an optimal prediction model, and taking a predicted value of flow data output by the optimal prediction model as the input of an ARED algorithm;
step 2.3: and carrying out smooth nonlinear processing on the packet discarding probability function of the ARED algorithm by utilizing a cubic curve function, and outputting the packet discarding probability through the ARED algorithm.
2. The method for managing the queue of the satellite network based on the traffic prediction as claimed in claim 1, wherein the step 2.1 comprises:
step 2.1.1: initializing an initial population in a differential evolution algorithm;
step 2.1.2: performing crossing, variation and selection on the initial population by using a differential evolution algorithm, and assigning the value corresponding to each generation of output individuals to a smoothing coefficient in a dynamic cubic exponential smoothing model after the maximum iteration times are reached;
step 2.1.3: taking the flow data with self-similarity as the input of a dynamic cubic exponential smoothing model, predicting the flow data by using the dynamic cubic exponential smoothing model, and outputting the predicted value of the flow data;
step 2.1.4: the sample mean square error f (Δ) is calculated using equation (1):
Figure FDA0003613114810000011
in the formula, xiIndicating the incoming flow data, SiA predicted value representing the output time queue, I represents the total number of flow data;
step 2.1.5: changing the value range of the initial population, and executing the step 2.1.2-step 2.1.4 again;
step 2.1.6: and (5) repeating the step 2.1.5, executing N times of calculation to obtain N sample mean square errors, and taking the individual value when the sample mean square error is minimum as the optimal value of the smoothing coefficient.
3. The method according to claim 1, wherein the step 2.3 of performing smooth nonlinear processing on the packet dropping probability function of the reed algorithm by using a cubic curve function includes:
step 2.3.1: the average queue length Qavg is calculated using equation (2):
Qavg＝(1-ωq)Qavg+ωq·q (2)
in the formula, omegaqRepresenting a weight value, q representing an initial queue length;
step 2.3.2: calculating the packet drop probability P using equation (3)b：
Figure FDA0003613114810000021
Where maxp represents the maximum packet drop probability, minthLower limit value, max, indicating a set queue thresholdthRepresents an upper limit value of the set queue threshold.
CN202110752131.5A 2021-06-30 2021-06-30 Satellite network queue management method based on flow prediction Active CN113472427B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110752131.5A CN113472427B (en) 2021-06-30 2021-06-30 Satellite network queue management method based on flow prediction

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110752131.5A CN113472427B (en) 2021-06-30 2021-06-30 Satellite network queue management method based on flow prediction

Publications (2)

Publication Number Publication Date
CN113472427A CN113472427A (en) 2021-10-01
CN113472427B true CN113472427B (en) 2022-07-15

Family

ID=77877797

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110752131.5A Active CN113472427B (en) 2021-06-30 2021-06-30 Satellite network queue management method based on flow prediction

Country Status (1)

Country Link
CN (1) CN113472427B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115334002B (en) * 2022-06-29 2024-04-26 沈阳理工大学 AOS intelligent frame generation method combining improved queue management algorithm under flow prediction

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104994031A (en) * 2015-07-13 2015-10-21 天津理工大学 Active queue self-adaptive management method ASRED
CN110636014A (en) * 2019-10-09 2019-12-31 大连大学 Satellite network queue scheduling algorithm based on traffic self-similarity characteristic

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104994031A (en) * 2015-07-13 2015-10-21 天津理工大学 Active queue self-adaptive management method ASRED
CN110636014A (en) * 2019-10-09 2019-12-31 大连大学 Satellite network queue scheduling algorithm based on traffic self-similarity characteristic

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
A Combined Forecasting Model for Satellite Network Self-Similar Traffic;Yuxia Bie et al;《IEEE Access》;20191003;第7卷;全文 *
Real-time Prediction of Docker Container Resource Load Based on A Hybrid Model of ARIMA and Triple Exponential Smoothing;Yulai Xie et al;《IEEE Transactions on Cloud Computing ( Early Access )》;20200422;全文 *
基于AOS 的Hurst-优先级自适应RED 与动态调度算法;别玉霞 等;《通信学报》;20121031;第33卷(第10期);第156-165页 *

Also Published As

Publication number Publication date
CN113472427A (en) 2021-10-01

Similar Documents

Publication Publication Date Title
CN113242568B (en) Task unloading and resource allocation method in uncertain network environment
CN110460880B (en) Industrial wireless streaming media self-adaptive transmission method based on particle swarm and neural network
CN113537514B (en) Digital twinning-based federal learning framework with high energy efficiency
CN109818775B (en) Short-term network flow prediction method
CN110233755B (en) Computing resource and frequency spectrum resource allocation method for fog computing in Internet of things
CN113259968A (en) Intelligent calculation method for power distribution network equipment based on information freshness
CN113472427B (en) Satellite network queue management method based on flow prediction
CN110519849B (en) Communication and computing resource joint allocation method for mobile edge computing
CN112105048B (en) Combined prediction method based on double-period Holt-Winters model and SARIMA model
CN113469426A (en) Photovoltaic output power prediction method and system based on improved BP neural network
CN112261725A (en) Data packet transmission intelligent decision method based on deep reinforcement learning
CN114169543A (en) Federal learning algorithm based on model obsolescence and user participation perception
CN109698925A (en) Real-time video jamming control method and device based on data-driven
CN114301910B (en) Cloud edge collaborative computing task unloading method in Internet of things environment
Bie et al. Queue management algorithm for satellite networks based on traffic prediction
CN116599860B (en) Network traffic gray prediction method based on reinforcement learning
CN113076663A (en) Dynamic hybrid precision model construction method and system
CN112862060B (en) Content caching method based on deep learning
CN110768827A (en) Task unloading method based on group intelligent algorithm
CN112600869B (en) Calculation unloading distribution method and device based on TD3 algorithm
WO2022121979A1 (en) Inner loop value adjustment method and device, storage medium, and electronic device
CN116305747A (en) Workflow multi-target scheduling method based on improved whale optimization algorithm
CN114449536A (en) 5G ultra-dense network multi-user access selection method based on deep reinforcement learning
CN114025321A (en) Massive power internet of things terminal access control method based on rapid uplink authorization
Liang et al. Stochastic Stackelberg Game Based Edge Service Selection for Massive IoT Networks

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