CN107682193A - A kind of common communication multi-business flow method of estimation - Google Patents
A kind of common communication multi-business flow method of estimation Download PDFInfo
- Publication number
- CN107682193A CN107682193A CN201710933527.3A CN201710933527A CN107682193A CN 107682193 A CN107682193 A CN 107682193A CN 201710933527 A CN201710933527 A CN 201710933527A CN 107682193 A CN107682193 A CN 107682193A
- Authority
- CN
- China
- Prior art keywords
- matrix
- mrow
- mtd
- msub
- mtr
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- 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
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Algebra (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Mathematical Physics (AREA)
- Probability & Statistics with Applications (AREA)
- Pure & Applied Mathematics (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
A kind of common communication multi-business flow method of estimation of the present invention, the relation between traffic matrix and link load is discussed, introduce constraints, build optimization object function;Then select appropriate object function to be iterated optimization, by introducing resolution matrix, and combine its attribute and good network chromatography priori result is obtained in iterative process;Then simplex method estimated flow matrix well is used.This method not only reduces the sensitiveness to apriority, and improve the speed of network chromatography without using statistical inference method estimated flow matrix.The present invention can realize that quick traffic matrix is accurately estimated and tracks its dynamic.
Description
Technical field
The present invention relates to the multi-business flow fast estimation technique field under common communication access network environment, more particularly to one
Kind common communication multi-business flow method of estimation.
Background technology
With the fast development of information technology, the mesh size of internet exponentially increases.In order to be set to network
Meter and planning, Virtual network operator need to know how data traffic is transmitted in a network, and traffic matrix reflects just
Data mobility status in network between all source and destination.Element in traffic matrix is referred to as source-destination (OD)
To (or stream).However, it is difficult to the traffic matrix in direct measurement network, and link load is readily available.Traffic matrix, route square
Battle array link load between relation be
Y=Ax
Wherein y represents link load (can be write as column vector), and x represents traffic matrix, and A represents route matrix, it every
One elements AijIf 1, mean that source-destination centering stream j by linking i;If 0, represent other situations.
Formula (1) shows given link load y and route matrix A, the solution with regard to that can find traffic matrix x.Due to
Route matrix A is typically to owe fixed and morbid state, and this is the Reverse Problem of a height morbid state.Network chromatography is that solve this problem
One of best method.Vardi (1996) describes network chromatography method to study this problem, and OD streams is modeled as first
Iid (independent, same to distribution) Poisson model.Cao etc. (2000) improves method and the OD streams of modeling are used as iid Gauss models.Yin
The problem of jade tablet etc. (2003) have studied network traffics Matrix Estimation, and make use of the Gravity Models of OD flows.They propose
Tomo-Gravity methods carry out estimated flow matrix, and obtain fairly accurate estimation.Zhang (2003) etc. uses regularization
Method estimated flow matrix, reaches more preferable effect.Based on Poisson model, Tebaldi and West (1998) use Bayes side
Method, it is still, very difficult due to calculating Posterior distrbutionp, therefore they are simulated using Markov chain-Monte-Carlo Simulation
They.As described in the paper of Medina etc. (2002), Vardi (1996), Cao (2000) et al. method traffic matrix
Apriority is very sensitive, and Zhang (2003) et al. method part reduces the sensitivity to apriority.
Certainly, the traffic matrix of large scale IP network can be obtained using the method for statistical inference.As Gunnar
(2004) described, for large scale IP traffic matrixs, what equation (1) represented is that a height owes fixed system, and this is due to net
Number of links in network is far smaller than the quantity of OD streams, and therefore, equation (1) has unlimited number of solution, how to find an average solution
The problem of being one extremely difficult.The method for solving this problem in the past is the additional constraint flowed by increase on OD, such as
Assuming that OD streams meet statistical distribution or they are modeled as into some typical models.But these methods are obtaining network traffics
The speed of Matrix Estimation is all very slow.
The content of the invention
In order to solve problem described in background technology, the present invention provides a kind of common communication multi-business flow method of estimation, this
Method not only reduces susceptibility of the traffic matrix estimation to apriority, and improves the speed of network chromatography.
In order to achieve the above object, the present invention is realized using following technical scheme:
A kind of common communication multi-business flow method of estimation, comprises the following steps:
Step 1: the relation between traffic matrix, route matrix and link load is analyzed first:
Y=Ax (1)
Wherein y represents link load (can be write as column vector), and x represents traffic matrix, and A represents route matrix, it every
One elements AijIf 1, mean that source-destination centering stream j by linking i;If 0, represent other situations;From formula
(1) in as can be seen that by building route matrix and measurement link load situation, the multiple services traffic matrix that communicates can be entered
Row estimation;
Step 2: introducing constraints, optimization object function is built;
Step 3: according to formula (1), it is assumed that the generalized inverse matrix of matrix A is A+, therefore traffic matrix can is expressed as
Formula (1), which is updated in formula (5), to be obtained
WhereinThe estimate of traffic matrix is meant that, x represents the actual value of traffic matrix;Therefore, if will be to flow square
Battle array is estimated, it is necessary to its generalized inverse matrix A is calculated according to route matrix A+;
Step 4: introduce resolution matrix R=A+A;When the element of every a line in matrix R is started with diagonal, two
Quickly reduce on individual opposite direction, and diagonal entry, close to 1, this shows estimate closer to actual value;Otherwise, it
Between have larger difference;Therefore, resolution matrix R represents the accuracy between estimate and true value, and only with priori
Information is related to estimating method, without related to value;Therefore, matrix A and its generalized inverse matrix A are passed through herein+Calculate resolution
Rate matrix R;
Step 5: by using searching method, selection matrix A U on the diagonal and line direction in resolution matrix R
Column vector, it is maximum linear independent subset, and corresponding to x U components;Therefore, equation (1) is decomposed into
Wherein V=N-U;
Step 6: the equation (7) in step 5, can obtain and carry out estimation institute using simplex method traffic matrix
The initial value x needed0;
Step 7: utilize initial value x0, according to the optimization object function of step 2, estimate is solved using simplex method
And output it.
The step 2 be specially:
Assuming that having n node and l link in a network, then N=n is there is in network2Individual OD streams, therefore formula
(1) the traffic matrix x cans in are expressed as x=(x1,x2,...,xN)T, link load y cans are expressed as y=(y1,
y2,...,yl), because each OD streams are non-negative, therefore each element in traffic matrix x meets xi>=0, according to
One appropriate optimization aim minf (x) of meaning selection of L1 norms=| x1|+|x2|+...+|xN|, due to constraints xi≥
0 limitation, therefore it is as follows to build object function:
Equation (2) is object function;Equation (3) and inequality (4) are the traffic matrix constraint function to be met.
Compared with prior art, the beneficial effects of the invention are as follows:
A kind of common communication multi-business flow method of estimation of the present invention, not only reduces the sensitiveness to apriority, and
Improve the speed of network chromatography.The present invention is not to use statistical inference method estimated flow matrix, but uses simplex method
Quick traffic matrix is accurately estimated.Present invention discusses the relation between traffic matrix and link load, and introduce
Constraints;Then appropriate object function is selected to be iterated optimization;Exist by introducing resolution matrix, and with reference to its attribute
Iterative process obtains good network chromatography priori result;Then simplex method estimated flow matrix well is used.This hair
It is bright to realize that quick traffic matrix is accurately estimated and tracks its dynamic.
Brief description of the drawings
Fig. 1 is the general flow chart of common communication multi-business flow method of estimation of the present invention;
Fig. 2 be one embodiment of the present invention in network A bilene the 106th, the 107th, No. 116 OD stream quick or
Valuation situation under situation of change at a slow speed, black is actual value;Grey is estimate;
Fig. 3 is that one embodiment of the present invention the 123rd, the 126th, No. 128 OD in network A bilene is flowed in measurement
Carve the valuation situation in the case of bigger or more outburst.Black is actual value;Grey is estimate;
Fig. 4 is amplification of the one embodiment of the present invention to Fig. 2 valuation situations between the 600th and 1000 time slots.Black
Be actual value;Grey is estimate;
Fig. 5 is amplification of the one embodiment of the present invention to Fig. 3 valuation situations between the 600th and 1000 time slots.Black
Be actual value;Grey is estimate;
Fig. 6 is that (x-axis is flow id for the space relative error of one embodiment of the present invention valuation and actual value;Stream according to
Average value from small to large progress order);
Fig. 7 is that (chronomere of x-axis is 5 points for the time relative error of one embodiment of the present invention valuation and actual value
Clock).
Embodiment
Embodiment provided by the invention is described in detail below in conjunction with accompanying drawing.
As shown in figure 1, a kind of common communication multi-business flow method of estimation, comprises the following steps:
Step 1: the relation between traffic matrix, route matrix and link load is analyzed first:
Y=Ax (1)
Wherein y represents link load (can be write as column vector), and x represents traffic matrix, and A represents route matrix, it every
One elements AijIf 1, mean that source-destination centering stream j by linking i;If 0, represent other situations;From formula
(1) in as can be seen that by building route matrix and measurement link load situation, the multiple services traffic matrix that communicates can be entered
Row estimation;
Step 2: introducing constraints, optimization object function is built;
Assuming that having n node and l link in a network, then N=n is there is in network2Individual OD streams, therefore formula
(1) the traffic matrix x cans in are expressed as x=(x1,x2,...,xN)T, link load y cans are expressed as y=(y1,
y2,...,yl), because each OD streams are non-negative, therefore each element in traffic matrix x meets xi>=0, according to
One appropriate optimization aim minf (x) of meaning selection of L1 norms=| x1|+|x2|+...+|xN|, due to constraints xi≥
0 limitation, therefore it is as follows to build object function:
Equation (2) is object function;Equation (3) and inequality (4) are the traffic matrix constraint function to be met.
Step 3: according to formula (1), it is assumed that the generalized inverse matrix of matrix A is A+, therefore traffic matrix can is expressed as
Formula (1), which is updated in formula (5), to be obtained
WhereinThe estimate of traffic matrix is meant that, x represents the actual value of traffic matrix;Therefore, if will be to flow square
Battle array is estimated, it is necessary to its generalized inverse matrix A+ is calculated according to route matrix A;
Step 4: introduce resolution matrix R=A+A;When the element of every a line in matrix R is started with diagonal, two
Quickly reduce on individual opposite direction, and diagonal entry, close to 1, this shows estimate closer to actual value;Otherwise, it
Between have larger difference;Therefore, resolution matrix R represents the accuracy between estimate and true value, and only with priori
Information is related to estimating method, without related to value;Therefore, resolution is calculated by matrix A and its generalized inverse matrix A+ herein
Rate matrix R;
Step 5: by using searching method, selection matrix A U on the diagonal and line direction in resolution matrix R
Column vector, it is maximum linear independent subset, and corresponding to x U components;Therefore, equation (1) is decomposed into
Wherein V=N-U;
Step 6: the equation (7) in step 5, can obtain and carry out estimation institute using simplex method traffic matrix
The initial value x needed0;
Step 7: utilize initial value x0, according to the equation (2) of step 2, (3) and inequality (4) optimization object function, adopt
Estimate is solved with simplex methodAnd output it.
Output result such as Fig. 2, Fig. 3 for quickly being estimated the route matrix of network using resolution ratio and simplex method,
Shown in Fig. 4, Fig. 5.Fig. 2 is that one embodiment of the present invention the 106th, the 107th, No. 116 OD in network A bilene is flowed fast
Speed or at a slow speed the valuation situation under situation of change.Black is actual value;Grey is estimate;Fig. 3 is of the invention a kind of real
Apply mode in network A bilene the 123rd, the 126th, No. 128 OD stream in the case of bigger or more the outburst of measurement moment
Valuation situation.Black is actual value;Grey is estimate;Fig. 4 is for one embodiment of the present invention to Fig. 2 in the 600th He
The amplification of valuation situation between 1000 time slots.Black is actual value;Grey is estimate;Fig. 5 is implemented for the present invention is a kind of
Amplification of the mode to Fig. 3 valuation situations between the 600th and 1000 time slots.Black is actual value;Grey is estimate.
In order to judge the traffic matrix method for quick estimating in the present invention, space relative error and time is incorporated herein
The concept of relative error, wherein space relative error formula expression are as follows
Wherein parameter n=1 ..., N, N are the sums of OD streams, and T is total time of measuring, | | | |2It is L2Norm.Fig. 5 makes
The space relative error of the inventive method is depicted with the real data in 1 to 7 March in 2004.X-axis is represented from minimum to most
Big orderly stream, sorts according to average value.Fig. 6 shows that, when OD streams are larger, estimation of the evaluated error less than smaller OD streams misses
Difference, therefore the present invention can more accurately estimate larger OD flows.
Time relative error formula expression is as follows
Wherein parameter t=1 ..., T.Fig. 7 depicts the time relative error of the present invention.X-axis is using the groove of 5 minutes to be single
Position.The present invention is included within the evaluated error of the inner peripheral flow amount estimation in 1 to 7 March in 2004.Fig. 7 display present invention's
Time relative error largely changes near 0.25.
The common communication multi-business flow method of estimation of the present invention uses RBF models, and the input of model is only using current
Link load Y (t)
The common communication multi-business flow method of estimation of the present invention employs estimated below mathematical modeling:
The present invention is trained using existing RBF networks, and is predicted, and is then iterated and is estimated with IPFP
Value.
Above example is implemented under premised on technical solution of the present invention, gives detailed embodiment and tool
The operating process of body, but protection scope of the present invention is not limited to the above embodiments.Method therefor is such as without spy in above-described embodiment
It is conventional method not mentionlet alone bright.
Claims (2)
1. a kind of common communication multi-business flow method of estimation, it is characterised in that comprise the following steps:
Step 1: the relation between traffic matrix, route matrix and link load is analyzed first:
Y=Ax (1)
Wherein y represents link load (can be write as column vector), and x represents traffic matrix, and A represents route matrix, it each
Elements AijIf 1, mean that source-destination centering stream j by linking i;If 0, represent other situations;From formula (1)
In as can be seen that by building route matrix and measurement link load situation, the multiple services traffic matrix of communicating can be carried out
Estimation;
Step 2: introducing constraints, optimization object function is built;
Step 3: according to route matrix A, generalized inverse matrix A is sought+:
According to formula (1), it is assumed that the generalized inverse matrix of matrix A is A+, therefore traffic matrix can is expressed as
<mrow>
<mover>
<mi>x</mi>
<mo>^</mo>
</mover>
<mo>=</mo>
<msup>
<mi>A</mi>
<mo>+</mo>
</msup>
<mi>y</mi>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>5</mn>
<mo>)</mo>
</mrow>
</mrow>
Formula (1), which is updated in formula (5), to be obtained
<mrow>
<mover>
<mi>x</mi>
<mo>^</mo>
</mover>
<mo>=</mo>
<msup>
<mi>A</mi>
<mo>+</mo>
</msup>
<mi>A</mi>
<mi>x</mi>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>6</mn>
<mo>)</mo>
</mrow>
</mrow>
WhereinThe estimate of traffic matrix is meant that, x represents the actual value of traffic matrix;Therefore, if wanting traffic matrix to enter
Row estimation, it is necessary to its generalized inverse matrix A is calculated according to route matrix A+;
Step 4: calculating resolution matrix R=A+A;It is opposite at two when the element of every a line in matrix R is started with diagonal
Direction on quickly reduce, and diagonal entry, close to 1, this shows estimate closer to actual value;Otherwise, between them
There is larger difference;Therefore, resolution matrix R represents the accuracy between estimate and true value, and only with prior information and
Estimating method is related, without related to value;Therefore, matrix A and its generalized inverse matrix A are passed through herein+Calculate resolution matrix
R;
Step 5: by using searching method on the diagonal and line direction in resolution matrix R, selection matrix A U arrange to
Amount, it is maximum linear independent subset, and corresponding to x U components;Therefore, equation (1) is decomposed into
<mrow>
<mi>y</mi>
<mo>=</mo>
<mo>&lsqb;</mo>
<msub>
<mi>A</mi>
<mi>U</mi>
</msub>
<mo>,</mo>
<msub>
<mi>A</mi>
<mi>V</mi>
</msub>
<mo>&rsqb;</mo>
<mfenced open = "[" close = "]">
<mtable>
<mtr>
<mtd>
<msub>
<mi>X</mi>
<mi>U</mi>
</msub>
</mtd>
</mtr>
<mtr>
<mtd>
<msub>
<mi>X</mi>
<mi>V</mi>
</msub>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>7</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein V=N-U;
Step 6: the equation (7) in step 5, can obtain using simplex method traffic matrix estimate it is required
Initial value x0;
Step 7: utilize initial value x0, according to the optimization object function of step 2, estimate is solved using simplex methodAnd will
It is exported.
2. a kind of common communication multi-business flow method of estimation according to claim 1, it is characterised in that the step 2
Specially:
Assuming that having n node and l link in a network, then N=n is there is in network2Individual OD streams, therefore in formula (1)
Traffic matrix x cans are expressed as x=(x1,x2,...,xN)T, link load y cans are expressed as y=(y1,y2,...,yl),
Because each OD streams are non-negative, therefore each element in traffic matrix x meets xi>=0, according to the meaning of L1 norms
One appropriate optimization aim minf (x) of justice selection=| x1|+|x2|+...+|xN|, due to constraints xi>=0 limitation, because
This structure object function is as follows:
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<mi>min</mi>
<mi> </mi>
<mi>f</mi>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<msub>
<mi>x</mi>
<mn>1</mn>
</msub>
<mo>+</mo>
<msub>
<mi>x</mi>
<mn>2</mn>
</msub>
<mo>+</mo>
<mo>...</mo>
<mo>+</mo>
<msub>
<mi>x</mi>
<mi>N</mi>
</msub>
</mrow>
</mtd>
<mtd>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<mi>y</mi>
<mo>=</mo>
<mi>A</mi>
<mi>x</mi>
</mrow>
</mtd>
<mtd>
<mrow>
<mo>(</mo>
<mn>3</mn>
<mo>)</mo>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mtable>
<mtr>
<mtd>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>&GreaterEqual;</mo>
<mn>0</mn>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
<mo>,</mo>
<mn>2</mn>
<mo>,</mo>
<mo>...</mo>
<mo>,</mo>
<mi>N</mi>
</mrow>
</mtd>
</mtr>
</mtable>
</mtd>
<mtd>
<mrow>
<mo>(</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
Equation (2) is object function;Equation (3) and inequality (4) are the traffic matrix constraint function to be met.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710933527.3A CN107682193A (en) | 2017-10-10 | 2017-10-10 | A kind of common communication multi-business flow method of estimation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710933527.3A CN107682193A (en) | 2017-10-10 | 2017-10-10 | A kind of common communication multi-business flow method of estimation |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107682193A true CN107682193A (en) | 2018-02-09 |
Family
ID=61137990
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710933527.3A Pending CN107682193A (en) | 2017-10-10 | 2017-10-10 | A kind of common communication multi-business flow method of estimation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107682193A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110909991A (en) * | 2019-11-08 | 2020-03-24 | 国网辽宁省电力有限公司电力科学研究院 | Rapid estimation device and method for optical cable fiber core remote intelligent scheduling service |
CN114978941A (en) * | 2022-05-24 | 2022-08-30 | 电子科技大学 | IPv6 network-oriented service flow measuring method |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6785240B1 (en) * | 2000-06-02 | 2004-08-31 | Lucent Technologies Inc. | Method for estimating the traffic matrix of a communication network |
CN102801629A (en) * | 2012-08-22 | 2012-11-28 | 电子科技大学 | Traffic matrix estimation method |
-
2017
- 2017-10-10 CN CN201710933527.3A patent/CN107682193A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6785240B1 (en) * | 2000-06-02 | 2004-08-31 | Lucent Technologies Inc. | Method for estimating the traffic matrix of a communication network |
CN102801629A (en) * | 2012-08-22 | 2012-11-28 | 电子科技大学 | Traffic matrix estimation method |
Non-Patent Citations (1)
Title |
---|
蒋定德: "大尺度IP流量矩阵估计关键技术研究", 《博士学位论文》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110909991A (en) * | 2019-11-08 | 2020-03-24 | 国网辽宁省电力有限公司电力科学研究院 | Rapid estimation device and method for optical cable fiber core remote intelligent scheduling service |
CN114978941A (en) * | 2022-05-24 | 2022-08-30 | 电子科技大学 | IPv6 network-oriented service flow measuring method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111130839B (en) | Flow demand matrix prediction method and system | |
Hsieh et al. | Learning neural PDE solvers with convergence guarantees | |
Kantas et al. | An overview of sequential Monte Carlo methods for parameter estimation in general state-space models | |
US8006220B2 (en) | Model-building optimization | |
Zhang et al. | Dynamic pricing, heterogeneous users and perception error: Probit-based bi-criterion dynamic stochastic user equilibrium assignment | |
Lin et al. | Gradient-free methods for deterministic and stochastic nonsmooth nonconvex optimization | |
CN102801629B (en) | Traffic matrix estimation method | |
CN106533778A (en) | Method for identifying key node of command and control network based on hierarchical flow betweenness | |
CN107682193A (en) | A kind of common communication multi-business flow method of estimation | |
Liu et al. | End-to-end learning of user equilibrium with implicit neural networks | |
Wong et al. | Graph neural network based surrogate model of physics simulations for geometry design | |
Park et al. | Mean value analysis of re-entrant line with batch machines and multi-class jobs | |
Chen et al. | Flow matching on general geometries | |
Li et al. | Infinite-fidelity coregionalization for physical simulation | |
Michau et al. | Estimating link-dependent origin-destination matrices from sample trajectories and traffic counts | |
CN109978138A (en) | The structural reliability methods of sampling based on deeply study | |
KR20190129422A (en) | Method and device for variational interference using neural network | |
CN115293367A (en) | Mixed federal learning method of scheduling model under small sample unbalanced data constraint | |
Wang et al. | Discrete stochastic optimization using linear interpolation | |
Zhu et al. | Identification of network sensor locations for estimation of traffic flow | |
Teknomo | Ideal relative flow distribution on directed network | |
CN102760182B (en) | Population travel demand calculating method based on compressive sensing technology | |
Telescu et al. | Stochastic Time-Domain Mapping for Comprehensive Uncertainty Assessment in Eye Diagrams | |
Jiang et al. | An accurate approach of large-scale IP traffic matrix estimation | |
Buchholz et al. | PH-graphs for analyzing shortest path problems with correlated traveling times |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180209 |
|
RJ01 | Rejection of invention patent application after publication |