CN107085942A - A kind of traffic flow forecasting method based on wolf pack algorithm, apparatus and system - Google Patents
A kind of traffic flow forecasting method based on wolf pack algorithm, apparatus and system Download PDFInfo
- Publication number
- CN107085942A CN107085942A CN201710495064.7A CN201710495064A CN107085942A CN 107085942 A CN107085942 A CN 107085942A CN 201710495064 A CN201710495064 A CN 201710495064A CN 107085942 A CN107085942 A CN 107085942A
- Authority
- CN
- China
- Prior art keywords
- wolf
- traffic flow
- neural network
- wavelet neural
- forecasting
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
-
- 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]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Strategic Management (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Human Resources & Organizations (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Game Theory and Decision Science (AREA)
- General Business, Economics & Management (AREA)
- Development Economics (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Traffic Control Systems (AREA)
Abstract
The embodiment of the invention discloses a kind of traffic flow forecasting method based on wolf pack algorithm, apparatus and system, including obtain traffic flow data;Traffic flow data progress is handled using the wavelet neural network forecasting traffic flow model pre-established and obtains forecasting traffic flow result;Wherein, wavelet neural network forecasting traffic flow model is formed based on wolf pack Algorithm for Training, and its training process is to calculate initialization wavelet neural network parameter according to historical data and wolf pack algorithm;Initialization wavelet neural network parameter is trained using wavelet neural network and historical data and obtains wavelet neural network forecasting traffic flow model.It can be seen that, the embodiment of the present invention improves predetermined speed and precision of prediction to a certain extent when the wavelet neural network forecasting traffic flow model gone out using the initialization wavelet neural network parameter training obtained based on wolf pack algorithm is predicted to traffic flow.
Description
Technical field
The present embodiments relate to road traffic technical field, more particularly to a kind of traffic flow based on wolf pack algorithm is pre-
Survey method, apparatus and system.
Background technology
When the traffic flow to road is predicted, it will usually by factors such as road conditions, time point, Changes in weather
Influence, so that cause road traffic flow data to have height uncertain, and rule is not obvious.In the prior art, to road
The network parameter of wavelet neural network is trained when the traffic flow on road is predicted using traditional Wavelet Neural Network Method, but
It is that the method used during due to using Traditional Wavelet neural metwork training network parameter is and basic BP neural network identical ladder
Descent method is spent, and gradient descent method has one-way, and the related network parameter of random generation, makes network parameter in optimization
During be extremely easily trapped into local minimum so that traffic flow predetermined speed and precision of prediction reduction.
Therefore, how a kind of traffic flow forecasting method based on wolf pack algorithm, device for solving above-mentioned technical problem is provided
And system turns into the problem of those skilled in the art needs to solve at present.
The content of the invention
The purpose of the embodiment of the present invention is to provide a kind of traffic flow forecasting method based on wolf pack algorithm, apparatus and system,
Predetermined speed and precision of prediction are being improved to a certain extent using process.
In order to solve the above technical problems, the embodiments of the invention provide a kind of forecasting traffic flow side based on wolf pack algorithm
Method, methods described includes:
Obtain traffic flow data;
The traffic flow data is handled using the wavelet neural network forecasting traffic flow model pre-established
Forecasting traffic flow result;Wherein, the wavelet neural network forecasting traffic flow model is formed based on wolf pack Algorithm for Training, its
Training process is:
Initialization wavelet neural network parameter is calculated according to historical data and wolf pack algorithm;
The initialization wavelet neural network parameter is trained using wavelet neural network and the historical data
Obtain the wavelet neural network forecasting traffic flow model.
Optionally, the process that initialization wavelet neural network parameter is calculated according to historical data and wolf pack algorithm
Specially:
Each network parameter is encoded to each individual wolf according to historical data, the position of each individual wolf with it is each
The network parameter is corresponded;
Foundation preset control parameters and corresponding strategy are found from individual wolf each described where head wolf after optimization
Position;
The position is carried out to decode and obtain network parameter corresponding with the position, and using the network parameter as first
Beginningization wavelet neural network parameter.
Optionally, the preset control parameters include the total quantity of individual wolf, maximum iteration, maximum migration number of times,
Predetermined number and pre-determined distance;
It is described to find the head wolf institute after optimization from individual wolf each described according to preset control parameters and corresponding strategy
The process of position be specially:
S2121:The wolf for choosing the predetermined number in addition to head wolf is changed as wolf, execution is visited by law of universal gravitation optimisation strategy
The migration behavior entered, and it regard the position corresponding to the maximum wolf of prey odorousness in current wolf pack as the migration side for visiting wolf
To;
S2122:The position for visiting wolf is updated, and until the prey odorousness for visiting wolf is dense more than the prey smell of the head wolf
When degree or current migration number of times reach maximum migration number of times, the corresponding position for visiting wolf is replaced to the position of the head wolf, it is described
Visiting wolf turns into new head wolf;
S2123:Violent wolf is called by the new head wolf, and using chaos intialization to the violent wolf variable of acquisition at
Reason, obtains initializing the position of violent wolf;
S2124:The position of violent wolf new individual is produced according to the position for initializing violent wolf, hunting for the violent wolf new individual is calculated
Thing odorousness, and when the prey odorousness is more than the prey odorousness of the new head wolf, the violent wolf is new
The position of individual replaces the position of the head wolf;Until the distance between violent wolf new individual and head wolf are less than pre-determined distance;
S2125:The jointly attack behavior of the adaptive policy optimization of inertia weight is performed, and the position of correct wolf is updated;
S2126:According to the position of " the victor is a king " Policy Updates head wolf, according still further to " powerhouse's existence " mechanism and " powerhouse's life
Deposit, the weak are the prey of the strong " principle colony renewal is carried out to wolf pack;
S2127:Judge whether reach that default precision or current iteration number of times are more than greatest iteration time between head wolf and prey
Number, if it is, using the head wolf as the head wolf after optimization, and export the position of the head wolf after the optimization;Otherwise, return
S2121。
Optionally, the process of the position for updating spy wolf is specially:
The position of the spy wolf is updated according to the first calculation relational expression;First calculation relational expression is:
Wherein, the x 'iRepresent that i-th is visited wolf more
Position after new;The xiRepresent the current location of i-th spy wolf, the xkRepresent that kth only visits the position of wolf, the rand (0,
1) uniformly distributed function of obedience 0 to 1, the x are representedbestRepresent the maximum position of current prey odorousness, the YikTable
Show that i-th is visited the prey odorousness function that wolf only visits wolf with respect to kth;
The YikObtained according to the second calculation relational expression, second calculation relational expression is:
Wherein, the G represents universal gravitational constant;Y (the xi) represent i-th prey odorousness for visiting wolf, the Y (xk) represent
Kth only visits the prey odorousness of wolf.
Optionally, the process of the position for producing violent wolf new individual according to the position for initializing violent wolf is specially:
The position of violent wolf new individual, the 3rd meter are obtained according to the position and the 3rd calculation relational expression for initializing violent wolf
Calculating relational expression is:
x′n=Yi+Ri(2xn,k- 1), wherein, the xn' be violent wolf new individual position, the xn,kRepresent kth time iteration
When violent wolf new individual position;The YiFor the odorousness of prey;The RiFor long-range raid zone radius;
The xn,kObtained according to the 4th calculation relational expression, the 4th calculation relational expression is xn,k+1=μ xn,k(1-xn,k),
Wherein, n ∈ [1, N], the N represents the total quantity of the individual wolf, and the μ represents the control parameter of chaos state, the k tables
Show iterations.
Optionally, the jointly attack behavior for performing the adaptive policy optimization of inertia weight, and the position of correct wolf is updated
Process be specially:
The jointly attack behavior of the adaptive policy optimization of inertia weight is performed, and is entered according to the position of the correct wolf of the 5th calculation relational expression
Row updates, and the 5th calculation relational expression isWherein, the MkRepresent iterations
Make the position of prey, the step for kcThe attack step-length of violent wolf is represented, the γ represents inertia weight;The γ is according to the 6th
Calculation relational expression is obtained, and the 6th calculation relational expression is:
Wherein, the γmaxRepresent maximum inertia weight;The γminRepresent most
Small inertia weight, the ZmaxThe maximum iteration is represented, the z represents current iteration number of times.
In order to solve the above technical problems, the embodiments of the invention provide a kind of forecasting traffic flow dress based on wolf pack algorithm
Put, described device includes:
Acquisition module, for obtaining traffic flow data;
Processing module, for using the wavelet neural network forecasting traffic flow model pre-established to the traffic flow data
Progress, which is handled, obtains forecasting traffic flow result;Wherein, the wavelet neural network forecasting traffic flow model includes:
Computing module, for calculating initialization wavelet neural network parameter according to historical data and wolf pack algorithm;
Training module, for using wavelet neural network and the historical data to the initialization wavelet neural network
Parameter, which is trained, obtains the wavelet neural network forecasting traffic flow model.
Optionally, the computing module includes:
Individual wolf coding unit, for each network parameter to be encoded into each individual wolf, Mei Gesuo according to historical data
The position and each network parameter for stating individual wolf are corresponded;
Head wolf seeks unit, excellent for being found according to preset control parameters and corresponding strategy from individual wolf each described
The position where head wolf after change;
Decoding unit, for the position decode to obtain network parameter corresponding with the position, and will be described
Network parameter is used as initialization wavelet neural network parameter.
In order to solve the above technical problems, the embodiments of the invention provide a kind of forecasting traffic flow system based on wolf pack algorithm
System, including the forecasting traffic flow device as described above based on wolf pack algorithm.
The embodiments of the invention provide a kind of traffic flow forecasting method based on wolf pack algorithm, apparatus and system, including:Obtain
Take traffic flow data;Traffic flow data is handled using the wavelet neural network forecasting traffic flow model pre-established
Forecasting traffic flow result;Wherein, wavelet neural network forecasting traffic flow model is formed based on wolf pack Algorithm for Training, and it is trained
Process is to calculate initialization wavelet neural network parameter according to historical data and wolf pack algorithm;Use wavelet neural network with
And historical data is trained to initialization wavelet neural network parameter and obtains wavelet neural network forecasting traffic flow model.
It can be seen that, the wavelet neural network forecasting traffic flow model that the embodiment of the present invention is used when being predicted to traffic flow
Initialization wavelet neural network parameter calculate and obtain according to historical data and wolf pack algorithm, because wolf pack algorithm has
The characteristics of search capability is strong, hunting zone is wide, therefore globally optimal solution can be largely converged on, therefore utilize and be based on wolf pack
The wavelet neural network forecasting traffic flow model that the initialization wavelet neural network parameter training that algorithm is obtained goes out is to traffic flow
When being predicted, predetermined speed and precision of prediction are improved to a certain extent.
Brief description of the drawings
Technical scheme in order to illustrate the embodiments of the present invention more clearly, below will be to institute in prior art and embodiment
The accompanying drawing needed to use is briefly described, it should be apparent that, drawings in the following description are only some implementations of the present invention
Example, for those of ordinary skill in the art, on the premise of not paying creative work, can also be obtained according to these accompanying drawings
Obtain other accompanying drawings.
Fig. 1 is a kind of schematic flow sheet of the traffic flow forecasting method based on wolf pack algorithm provided in an embodiment of the present invention;
The freeway traffic for the traffic flow forecasting method based on wolf pack algorithm that Fig. 2 is provided using the embodiment of the present invention
Flow predictive simulation schematic diagram;
Fig. 3 is the freeway traffic flow prediction using the traffic flow forecasting method of wavelet neural network of the prior art
Emulate schematic diagram;
Fig. 4 is a kind of structural representation of the forecasting traffic flow device based on wolf pack algorithm provided in an embodiment of the present invention;
Fig. 5 is a kind of structural representation of wavelet neural network forecasting traffic flow model provided in an embodiment of the present invention.
Embodiment
The embodiments of the invention provide a kind of traffic flow forecasting method based on wolf pack algorithm, apparatus and system, using
Process improves predetermined speed and precision of prediction to a certain extent.
To make the purpose, technical scheme and advantage of the embodiment of the present invention clearer, below in conjunction with the embodiment of the present invention
In accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described, it is clear that described embodiment is
A part of embodiment of the present invention, rather than whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art
The every other embodiment obtained under the premise of creative work is not made, belongs to the scope of protection of the invention.
Fig. 1 is refer to, Fig. 1 is a kind of stream of the traffic flow forecasting method based on wolf pack algorithm provided in an embodiment of the present invention
Journey schematic diagram.This method includes:
S11:Obtain traffic flow data;
S12:Traffic flow data is handled using the wavelet neural network forecasting traffic flow model pre-established
Forecasting traffic flow result;Wherein, wavelet neural network forecasting traffic flow model is formed based on wolf pack Algorithm for Training, and it is trained
Process is:
S21:Initialization wavelet neural network parameter is calculated according to historical data and wolf pack algorithm;
S22:Initialization wavelet neural network parameter is trained using wavelet neural network and historical data and obtained
Wavelet neural network forecasting traffic flow model.
It should be noted that the traffic flow data of road (such as highway) is obtained, it is logical according to these traffic flow datas
Cross the wavelet neural network forecasting traffic flow model pre-established and carry out forecasting traffic flow.It is used for training in the embodiment of the present invention
The initialization wavelet neural network parameter of wavelet neural network forecasting traffic flow model is (for example, wavelet neural network connection weight
And threshold parameter etc.) calculated by wolf pack algorithm, gone through for example, can be obtained in advance from traffic data control centre
History data (namely historical traffic flow data), and calculating processing is carried out using wolf pack algorithm, so as to obtain initializing wavelet neural
Network parameter, recycles and historical data and initialization wavelet neural network parameter is inputted and be trained into wavelet neural network
Obtain wavelet neural network forecasting traffic flow model.
Specifically, in actual applications, for example for certain section of highway traffic flow prediction, can first from this at a high speed
Traffic flow data is obtained in the database of the corresponding traffic data control centre of highway, it is possible to will choose prediction section 2016
31 days Mays, totally 2976 traffic flow datas were used as experiment data.In order that it is more accurate to predict the outcome, it will can also obtain
Original traffic flow data carry out data processing and include Noise reducing of data, will after disorder data recognition and reparation and normalized
A portion traffic flow data (for example, using 24 days before the middle of the month totally 2016 traffic flow datas) is as historical data, by this
Partial history data are instructed by being used as training sample after phase space reconfiguration to these historical datas using wolf pack algorithm
Practice, obtain initializing wavelet neural network parameter, another part data (672 traffic flow datas in i.e. last 7 days) are entered
Test sample is used as after row phase space reconfiguration (i.e. as the traffic flow data for predicting).It is, being gone through using the first 24 days
History data training initialization wavelet neural network parameter builds wavelet neural network forecasting traffic flow model, then good by component
Wavelet neural network forecasting traffic flow model carried out single-point Single-step Prediction to the magnitude of traffic flow of latter 7 days, to be predicted the outcome.
Certainly, above-mentioned to be merely illustrative, historical data and prediction data can be gone through using same group in actual applications
History traffic flow data or different historical traffic flow datas, are specifically used as historical data using which traffic flow data
With prediction data can be according to actual conditions depending on, the embodiment of the present invention does not do special restriction to this, can realize of the invention real
Apply the purpose of example.
The embodiments of the invention provide a kind of traffic flow forecasting method based on wolf pack algorithm, including:Obtain traffic fluxion
According to;Traffic flow data progress is handled using the wavelet neural network forecasting traffic flow model pre-established and obtains forecasting traffic flow
As a result;Wherein, wavelet neural network forecasting traffic flow model is formed based on wolf pack Algorithm for Training, and its training process is foundation
Historical data and wolf pack algorithm calculate initialization wavelet neural network parameter;Using wavelet neural network and historical data
Initialization wavelet neural network parameter is trained and obtains wavelet neural network forecasting traffic flow model.
It can be seen that, the wavelet neural network forecasting traffic flow model that the embodiment of the present invention is used when being predicted to traffic flow
Initialization wavelet neural network parameter calculate and obtain according to historical data and wolf pack algorithm, because wolf pack algorithm has
The characteristics of search capability is strong, hunting zone is wide, therefore globally optimal solution can be largely converged on, therefore utilize and be based on wolf pack
The wavelet neural network forecasting traffic flow model that the initialization wavelet neural network parameter training that algorithm is obtained goes out is to traffic flow
When being predicted, predetermined speed and precision of prediction are improved to a certain extent.
The embodiment of the invention discloses a kind of traffic flow forecasting method based on wolf pack algorithm, relative to a upper embodiment,
The present embodiment has made further instruction and optimization to technical scheme.
It should be noted that being inputted before training wavelet neural network forecasting traffic flow model, it is necessary to pre-set
Traffic flow data or historical data length and the control parameter of model.For example, input length can be set into TFS (just
Integer), then input length be TFS traffic flow data (or Traffic Flow Time Series) be c=c (i) | i=1,2, TFS;
Set control parameter can include:Input, i.e. input layer number;Hidden, i.e. wavelet layer nerve
First number;Ouput, i.e. output layer neuron number, wherein, Input≤TFS.
In addition it is also necessary to set up wavelet neural network freeway traffic flow forecast model:
Wherein, o represents that wavelet neural network traffic flow is exported;wijRepresent i-th of input of connection and the company of j-th of Wavelet Element
Connect weights;(F (1), F (2), F (Input)) is the traffic flow data (i.e. phase space reconfiguration traffic flow input data) that is inputted;
vijRepresent the weights of connection wavelet layer and output layer;bjRepresent j-th of translation coefficient;ajRepresent j-th of coefficient of dilatation;L is represented
Wavelet basis function, andWherein, t is chronomere second.In the embodiment of the present invention, by calculating
Go out to initialize wavelet neural network parameter wij, vij, aj, bj(i=1,2 ..., Input;J=1,2 ..., Output), you can enter
One step obtains wavelet neural network forecasting traffic flow model, and for the prediction to traffic flow.Specifically:
In the S21 of a upper embodiment, initialization wavelet neural network is calculated according to historical data and wolf pack algorithm
The process of parameter, is specifically as follows:
S211:Each network parameter is encoded to each individual wolf according to historical data, the position of each individual wolf with it is every
Individual network parameter is corresponded;
Specifically, the individual wolf of coding is by each network parameter wij, vij, aj, bj(i=1,2 ..., Input;J=1,
2 ..., Output) encoded, wherein, the position of i-th of individual wolf of D dimensions is xi=(wij,vij,aj,bj)T, i.e. each individual
The position of wolf is corresponded with corresponding network parameter, and it is wavelet neural network number of parameters summation that D, which represents,.
S212:Foundation preset control parameters and corresponding strategy are found from each individual wolf where the head wolf after optimization
Position;
It should be noted that being controlled the setting of parameter in advance, each preset control parameters, preset control parameters are obtained
It can include, total quantity N (namely visiting the total number of wolf, violent wolf and head wolf), the maximum iteration Z of individual wolfmax, it is maximum
Migration number of times Tmax, predetermined number h (visit wolf number) and the pre-determined distance (minimum range between i.e. violent wolf and head wolf
Hnear).Current migration number of times can also be represented with T;Use YiRepresent that (namely current wolf detects for the prey odorousness of current wolf
Prey odorousness);Use YleadRepresent the prey odorousness of head wolf;Current algebraically iterations is represented with z.
Then, the head after optimization is found from each individual wolf according to preset control parameters and corresponding strategy in above-mentioned S212
The process of position where wolf, is specifically as follows:
It should be noted that firstly the need of the migration strategy for determining wolf pack, that is, choosing the wolf of the predetermined number in addition to head wolf
As wolf is visited, perform by the improved migration behavior of law of universal gravitation optimisation strategy.Specific such as S2121-S2122:
S2121:The wolf for choosing the predetermined number in addition to head wolf is changed as wolf, execution is visited by law of universal gravitation optimisation strategy
The migration behavior entered, and it regard the position corresponding to the maximum wolf of prey odorousness in current wolf pack as the migration side for visiting wolf
To;S2122:The position for visiting wolf is updated, and until the prey odorousness for visiting wolf is more than prey odorousness (the i.e. Y of head wolfi>
Ylead) or currently migration number of times T reaches maximum migration number of times TmaxWhen, the corresponding position for visiting wolf is replaced to the position of head wolf, visited
Wolf turns into new head wolf;
Optionally, the preset control parameters in the embodiment of the present invention can include the total quantity of individual wolf, greatest iteration time
Number, maximum migration number of times, predetermined number and pre-determined distance;
Certainly, preset control parameters are not limited only to include above-mentioned several parameters, can also include other parameters, specifically
Depending on can be according to actual conditions, the embodiment of the present invention do special restriction to this, can realize the purpose of the embodiment of the present invention
.
Further, the process for the position for visiting wolf is updated in S2122, is specifically as follows:
The position for visiting wolf is updated according to the first calculation relational expression;First calculation relational expression is:
Wherein, x 'iRepresent that i-th is visited after wolf renewal
Position;xiRepresent the current location of i-th spy wolf, xkRepresent that kth only visits the position of wolf, rand (0,1) represents to obey 0 to 1
Uniformly distributed function, xbestRepresent the maximum position of current prey odorousness, YikRepresent that visit wolf for i-th only visits wolf with respect to kth
Prey odorousness function;
YikObtained according to the second calculation relational expression, the second calculation relational expression is:
Wherein, G represents universal gravitational constant;Y(xi) represent the prey odorousness of i-th spy wolf (that is, wavelet neural network is trained
The error of output valve and its desired value is target function value), Y (xk) represent kth, only visit the prey odorousness of wolf.
It should be noted that above-mentioned Y (xi) can be by target function typeObtain, and wolf
The average value that wolf position is visited in group is metFormula.
After the migration strategy of wolf pack is determined, then it needs to be determined that the long-range raid strategy of wolf pack, by top produce it is new
Head wolf call violent wolf, violent wolf is performed chaos policy optimization long-range raid behavior, it is specific such as S2123-S2124:
S2123:Violent wolf is called by new head wolf, and the violent wolf variable of acquisition handled using chaos intialization,
Obtain initializing the position of violent wolf;
S2124:The position of violent wolf new individual is produced according to the position for initializing violent wolf, the prey gas of violent wolf new individual is calculated
Taste concentration, and when prey odorousness is more than the prey odorousness of new head wolf, the position of violent wolf new individual is replaced into head
The position of wolf;Until the distance between violent wolf new individual and head wolf are less than pre-determined distance;
Further, the process of the position of violent wolf new individual is produced according to the position for initializing violent wolf in S2124, specifically
Can be:
The position of violent wolf new individual is obtained according to the position and the 3rd calculation relational expression for initializing violent wolf, the 3rd calculates pass
It is that formula is:
x′n=Yi+Ri(2xn,k- 1), wherein, xn' be violent wolf new individual position, xn,kRepresent that violent wolf is new during kth time iteration
The position Y of individualiFor the odorousness of prey;RiFor long-range raid zone radius;
Wherein, xn,kIt can be obtained according to the 4th calculation relational expression, the 4th calculation relational expression is xn,k+1=μ xn,k(1-xn,k),
Wherein, n ∈ [1, N], N represents the total quantity of individual wolf, and μ represents the control parameter of chaos state, and k represents iterations.
Specifically, the process is that is, by variable xnChaotic maps are to optimized variable x 'n, x 'nTo be changed into the wolf of a wolf
Prey concentration smell Y at placeiCentered on, with RiOn the long-range raid region of radius, to calculate prey odorousness Y (x 'n), and
And update the maximum prey odorousness Y during chaos iterationbest(x′n), if the prey odorousness of current wolf is more than
The prey odorousness of head wolf, then replace head wolf.Until the distance between violent wolf and head wolf are less than pre-determined distance Hnear。
It should be noted that the value of the μ in the embodiment of the present invention can be 4, can be completely into chaos when μ takes 4
State.Certainly, in actual applications, μ value is not limited only to take 4, or other numerical value, its concrete numerical value can root
Depending on actual conditions, the embodiment of the present invention does not do special restriction to this, can realize the purpose of the embodiment of the present invention.
After the long-range raid strategy of wolf pack is determined, then it needs to be determined that the jointly attack strategy of wolf pack, performs inertia weight adaptive
The jointly attack behavior of policy optimization, it is specific such as S2125:
S2125:The jointly attack behavior of the adaptive policy optimization of inertia weight is performed, and the position of correct wolf is updated;
Further, the jointly attack behavior of the adaptive policy optimization of inertia weight is performed in S2125, and the position of correct wolf is entered
The process that row updates, is specifically as follows:
The jointly attack behavior of the adaptive policy optimization of inertia weight is performed, and is entered according to the position of the correct wolf of the 5th calculation relational expression
Row updates, and the 5th calculation relational expression isWherein, MkIt is that k makes prey to represent iterations
Position, stepcThe attack step-length of violent wolf is represented, γ represents inertia weight;γ is obtained according to the 6th calculation relational expression, the 6th meter
Calculating relational expression is:
Wherein, γmaxRepresent maximum inertia weight;γminRepresent minimum inertia
Weight, ZmaxMaximum iteration is represented, z represents current iteration number of times.
S2126:According to the position of " the victor is a king " Policy Updates head wolf, according still further to " powerhouse's existence " mechanism and " powerhouse's life
Deposit, the weak are the prey of the strong " principle colony renewal is carried out to wolf pack;
Specifically, when carrying out colony's renewal to wolf pack, the P wolf farthest apart from prey (target function value) can be washed in a pan
Eliminate and randomly generate P wolf, to be updated to wolf pack.Wherein, P ∈ [N/ (2 β), N/ β], β are renewal scale factor.
S2127:Judge whether reach that default precision or current iteration number of times are more than greatest iteration time between head wolf and prey
Number, if it is, using head wolf as the head wolf after optimization, and export the position of the head wolf after optimization;Otherwise, S2121 is returned.
S213:Position decode to obtain network parameter corresponding with position, and network parameter is small as initializing
Ripple neural network parameter.
Specifically, because the position of each individual wolf in wolf pack is corresponded with network parameter, so after optimization is found
Head wolf after, the position of the head wolf after optimization, which is decoded, can obtain initializing wavelet neural network parameter.
It should be noted that calculate initialization wavelet neural network parameter after, can use wavelet neural network with
And historical data is trained to initialization wavelet neural network parameter and obtains wavelet neural network forecasting traffic flow model.I.e.
S22 process is specific as follows:
S221:According to Input input layer, (inputted using G-P algorithms reconstruct traffic flow sequence phase space
Input historical datas (i.e. traffic flow sequence) prediction Input+1 Traffic Flow Time Series) obtain training input sample and instruction
Practice output sample.
S222:Set up training objective functionWherein E represents wavelet neural network
The mean square error function of forecasting traffic flow desired value and network real output value;Sp represents training sample group number;sjRepresent j-th
Traffic flow desired value is exported.
S223:If | E | more than setting value, such as according to formula AndWavelet neural network parameter is updated, S222 is skipped to, to correct wavelet neural network ginseng
Number;Wherein η is wavelet neural network Studying factors.
The wavelet neural network parameter w trainedij, vij, aj, bj(i=1,2 ..., Input;J=1,2 ...,
Output prediction wavelet neural network) is substituted into, wavelet neural network forecasting traffic flow model can be obtained, and utilize the friendship obtained
Through-flow data and calculation relational expressionObtain wavelet neural network prediction output.
In addition, Fig. 2 and Fig. 3 are refer to, the traffic flow based on wolf pack algorithm that Fig. 2 is provided for the use embodiment of the present invention
The freeway traffic flow predictive simulation schematic diagram of Forecasting Methodology, Fig. 3 is the friendship using wavelet neural network of the prior art
The freeway traffic flow predictive simulation schematic diagram of through-flow Forecasting Methodology.IWN-WNN in Fig. 2 is represented based on the small of wolf pack algorithm
Ripple neural net prediction method;WNN in Fig. 3 represents the Forecasting Methodology based on wavelet neural network.From Fig. 2 and Fig. 3, sheet
The accuracy for the traffic flow forecasting method based on wolf pack algorithm that inventive embodiments are provided is higher, and prediction effect is more preferable.
Accordingly, the embodiment of the invention discloses a kind of forecasting traffic flow device based on wolf pack algorithm.Specifically it refer to
Fig. 4 Fig. 4 is a kind of structural representation of the forecasting traffic flow device based on wolf pack algorithm provided in an embodiment of the present invention.Above-mentioned
On the basis of embodiment:
The device includes:
Acquisition module 1, for obtaining traffic flow data;
Processing module 2, for being entered using the wavelet neural network forecasting traffic flow model pre-established to traffic flow data
Row processing obtains forecasting traffic flow result;Wherein, wavelet neural network forecasting traffic flow model includes:
Computing module, for calculating initialization wavelet neural network parameter according to historical data and wolf pack algorithm;
Training module, for being carried out using wavelet neural network and historical data to initialization wavelet neural network parameter
Training obtains wavelet neural network forecasting traffic flow model.
It should be noted that a kind of system for forecasting traffic flow based on wolf pack algorithm provided in an embodiment of the present invention, makes
Predetermined speed and precision of prediction can be improved to a certain extent during.
In addition, for specific Jie of the traffic flow forecasting method based on wolf pack algorithm involved in the embodiment of the present invention
Continue, refer to above method embodiment, the application will not be repeated here.
On the basis of above-described embodiment, Fig. 5 is refer to, Fig. 5 is a kind of Wavelet Neural Network provided in an embodiment of the present invention
The structural representation of network forecasting traffic flow model.
Optionally, computing module includes:
Individual wolf coding unit, for each network parameter to be encoded into each individual wolf according to historical data, per each and every one
The position of body wolf is corresponded with each network parameter;
Head wolf seeks unit, for being found according to preset control parameters and corresponding strategy from each individual wolf after optimization
Head wolf where position;
Decoding unit, for position to be carried out to decode and obtain network parameter corresponding with position, and using network parameter as
Initialize wavelet neural network parameter.
In order to solve the above technical problems, the embodiments of the invention provide a kind of forecasting traffic flow system based on wolf pack algorithm
System, including the forecasting traffic flow device based on wolf pack algorithm described above.
It should be noted that a kind of system for forecasting traffic flow based on wolf pack algorithm provided in an embodiment of the present invention, makes
Predetermined speed and precision of prediction can be improved to a certain extent during.
In addition, for specific Jie of the traffic flow forecasting method based on wolf pack algorithm involved in the embodiment of the present invention
Continue, refer to above method embodiment, the application will not be repeated here.
It should also be noted that, in this manual, such as first and second or the like relational terms be used merely to by
One entity or operation make a distinction with another entity or operation, and not necessarily require or imply these entities or operation
Between there is any this actual relation or order.Moreover, term " comprising ", "comprising" or its any other variant meaning
Covering including for nonexcludability, so that process, method, article or equipment including a series of key elements not only include that
A little key elements, but also other key elements including being not expressly set out, or also include be this process, method, article or
The intrinsic key element of equipment.In the absence of more restrictions, the key element limited by sentence "including a ...", is not arranged
Except also there is other identical element in the process including the key element, method, article or equipment.
The foregoing description of the disclosed embodiments, enables professional and technical personnel in the field to realize or using the present invention.
A variety of modifications to these embodiments will be apparent for those skilled in the art, as defined herein
General Principle can be realized in other embodiments without departing from the spirit or scope of the present invention.Therefore, it is of the invention
The embodiments shown herein is not intended to be limited to, and is to fit to and principles disclosed herein and features of novelty phase one
The most wide scope caused.
Claims (9)
1. a kind of traffic flow forecasting method based on wolf pack algorithm, it is characterised in that methods described includes:
Obtain traffic flow data;
The traffic flow data progress is handled using the wavelet neural network forecasting traffic flow model pre-established and obtains traffic
Stream predicts the outcome;Wherein, the wavelet neural network forecasting traffic flow model is formed based on wolf pack Algorithm for Training, and it is trained
Process is:
Initialization wavelet neural network parameter is calculated according to historical data and wolf pack algorithm;
The initialization wavelet neural network parameter is trained using wavelet neural network and the historical data and obtained
The wavelet neural network forecasting traffic flow model.
2. the traffic flow forecasting method according to claim 1 based on wolf pack algorithm, it is characterised in that described according to history
Data and wolf pack algorithm calculate initialization wavelet neural network parameter process be specially:
Each network parameter is encoded to each individual wolf according to historical data, the position of each individual wolf with it is each described
Network parameter is corresponded;
The position where the head wolf after optimization is found from individual wolf each described according to preset control parameters and corresponding strategy;
The position decode to obtain network parameter corresponding with the position, and regard the network parameter as initialization
Wavelet neural network parameter.
3. the traffic flow forecasting method according to claim 2 based on wolf pack algorithm, it is characterised in that
The preset control parameters include the total quantity of individual wolf, maximum iteration, maximum migration number of times, predetermined number and
Pre-determined distance;
The foundation preset control parameters and corresponding strategy are found from individual wolf each described where head wolf after optimization
The process of position is specially:
S2121:The wolf of the predetermined number in addition to head wolf is chosen as wolf is visited, is performed improved by law of universal gravitation optimisation strategy
Migration behavior, and it regard the position corresponding to the maximum wolf of prey odorousness in current wolf pack as the migration direction for visiting wolf;
S2122:Update visit wolf position, and until visit wolf prey odorousness be more than the head wolf prey odorousness or
When current migration number of times reaches maximum migration number of times, the corresponding position for visiting wolf is replaced to the position of the head wolf, the spy wolf
As new head wolf;
S2123:Violent wolf is called by the new head wolf, and the violent wolf variable of acquisition handled using chaos intialization,
Obtain initializing the position of violent wolf;
S2124:The position of violent wolf new individual is produced according to the position for initializing violent wolf, the prey gas of the violent wolf new individual is calculated
Taste concentration, and when the prey odorousness is more than the prey odorousness of the new head wolf, by the violent wolf new individual
Position replace the head wolf position;Until the distance between violent wolf new individual and head wolf are less than pre-determined distance;
S2125:The jointly attack behavior of the adaptive policy optimization of inertia weight is performed, and the position of correct wolf is updated;
S2126:According to the position of " the victor is a king " Policy Updates head wolf, according still further to " powerhouse's existence " mechanism and " powerhouse survives, weak
Meat is eaten by force " principle colony renewal is carried out to wolf pack;
S2127:Judge whether reach that default precision or current iteration number of times are more than maximum iteration between head wolf and prey, such as
Fruit is then using the head wolf as the head wolf after optimization, and to export the position of the head wolf after the optimization;Otherwise, return
S2121。
4. the traffic flow forecasting method according to claim 3 based on wolf pack algorithm, it is characterised in that wolf is visited in the renewal
The process of position be specially:
The position of the spy wolf is updated according to the first calculation relational expression;First calculation relational expression is:
Wherein, the x 'iRepresent that i-th is visited after wolf renewal
Position;The xiRepresent the current location of i-th spy wolf, the xkRepresent that kth only visits the position of wolf, rand (0, the 1) table
Show the uniformly distributed function of obedience 0 to 1, the xbestRepresent the maximum position of current prey odorousness, the YikRepresent i-th
Only visit the prey odorousness function that wolf only visits wolf with respect to kth;
The YikObtained according to the second calculation relational expression, second calculation relational expression is:
Wherein, the G represents universal gravitational constant;Y (the xi) represent i-th prey odorousness for visiting wolf, the Y (xk) represent
Kth only visits the prey odorousness of wolf.
5. the traffic flow forecasting method according to claim 3 based on wolf pack algorithm, it is characterised in that the foundation is initial
The process for the position that the position for changing violent wolf produces violent wolf new individual is specially:
The position of violent wolf new individual is obtained according to the position and the 3rd calculation relational expression for initializing violent wolf, the described 3rd calculates pass
It is that formula is:
xn'=Yi+Ri(2xn,k- 1), wherein, the x 'nFor the position of violent wolf new individual, the xn,kRepresent violent during kth time iteration
The position of wolf new individual;The YiFor the odorousness of prey;The RiFor long-range raid zone radius;
The xn,kObtained according to the 4th calculation relational expression, the 4th calculation relational expression is xn,k+1=μ xn,k(1-xn,k), wherein,
N ∈ [1, N], the N represent the total quantity of the individual wolf, and the μ represents the control parameter of chaos state, and the k represents to change
Generation number.
6. the traffic flow forecasting method according to claim 3 based on wolf pack algorithm, it is characterised in that the execution inertia
The jointly attack behavior of the adaptive policy optimization of weight, and the process that is updated of position of correct wolf is specially:
The jointly attack behavior of the adaptive policy optimization of inertia weight is performed, and is carried out more according to the position of the correct wolf of the 5th calculation relational expression
Newly, the 5th calculation relational expression isWherein, the MkRepresent that iterations makes for k
The position of prey, the stepcThe attack step-length of violent wolf is represented, the γ represents inertia weight;The γ is calculated according to the 6th
Relational expression is obtained, and the 6th calculation relational expression is:
Wherein, the γmaxRepresent maximum inertia weight;The γminRepresent minimum
Inertia weight, the ZmaxThe maximum iteration is represented, the z represents current iteration number of times.
7. a kind of forecasting traffic flow device based on wolf pack algorithm, it is characterised in that described device includes:
Acquisition module, for obtaining traffic flow data;
Processing module, for being carried out using the wavelet neural network forecasting traffic flow model pre-established to the traffic flow data
Processing obtains forecasting traffic flow result;Wherein, the wavelet neural network forecasting traffic flow model includes:
Computing module, for calculating initialization wavelet neural network parameter according to historical data and wolf pack algorithm;
Training module, for initializing wavelet neural network parameter to described using wavelet neural network and the historical data
It is trained and obtains the wavelet neural network forecasting traffic flow model.
8. the forecasting traffic flow device according to claim 7 based on wolf pack algorithm, it is characterised in that the computing module
Including:
Individual wolf coding unit, for each network parameter to be encoded into each individual wolf according to historical data, each described
The position of body wolf is corresponded with each network parameter;
Head wolf seeks unit, for according to preset control parameters and corresponding strategy from being found each described in individual wolf after optimization
Head wolf where position;
Decoding unit, for carrying out the position to decode and obtain network parameter corresponding with the position, and by the network
Parameter is used as initialization wavelet neural network parameter.
9. a kind of system for forecasting traffic flow based on wolf pack algorithm, it is characterised in that including base as claimed in claim 7 or 8
In the forecasting traffic flow device of wolf pack algorithm.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710495064.7A CN107085942B (en) | 2017-06-26 | 2017-06-26 | Traffic flow prediction method, device and system based on wolf colony algorithm |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710495064.7A CN107085942B (en) | 2017-06-26 | 2017-06-26 | Traffic flow prediction method, device and system based on wolf colony algorithm |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107085942A true CN107085942A (en) | 2017-08-22 |
CN107085942B CN107085942B (en) | 2021-01-26 |
Family
ID=59607222
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710495064.7A Active CN107085942B (en) | 2017-06-26 | 2017-06-26 | Traffic flow prediction method, device and system based on wolf colony algorithm |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107085942B (en) |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107688702A (en) * | 2017-08-24 | 2018-02-13 | 西安理工大学 | A kind of river flood traffic camouflaging rule analogy method based on wolf pack algorithm |
CN107917734A (en) * | 2017-11-29 | 2018-04-17 | 国网吉林省电力有限公司信息通信公司 | Cable's Fault Forecasting Methodology based on temperature and resistance |
CN108422997A (en) * | 2018-05-15 | 2018-08-21 | 南京航空航天大学 | A kind of automobile active safety cooperative control system and method based on wolf pack algorithm |
CN108491923A (en) * | 2018-04-10 | 2018-09-04 | 吉林大学 | Based on the pest image-recognizing method for improving wolf pack algorithm optimization Bayesian network |
CN108898223A (en) * | 2018-07-11 | 2018-11-27 | 国家海洋技术中心 | A kind of ocean observation device operating status method for detecting abnormality and device |
CN108918137A (en) * | 2018-06-08 | 2018-11-30 | 华北水利水电大学 | Fault Diagnosis of Gear Case devices and methods therefor based on improved WPA-BP neural network |
CN109347900A (en) * | 2018-08-23 | 2019-02-15 | 东华理工大学 | Based on the cloud service system Adaptive evolution method for improving wolf pack algorithm |
CN110444022A (en) * | 2019-08-15 | 2019-11-12 | 平安科技(深圳)有限公司 | The construction method and device of traffic flow data analysis model |
CN110824293A (en) * | 2019-10-15 | 2020-02-21 | 淮阴工学院 | Power grid fault diagnosis method based on multi-feature fusion parameters of wolf pack algorithm |
CN111259498A (en) * | 2020-01-14 | 2020-06-09 | 重庆大学 | Axle system thermal error modeling method and thermal error compensation system based on LSTM neural network |
CN111768622A (en) * | 2020-06-23 | 2020-10-13 | 南通大学 | Short-time traffic prediction method based on improved wolf algorithm |
CN111964574A (en) * | 2020-06-23 | 2020-11-20 | 安徽理工大学 | Three-dimensional laser target sphere center fitting method based on weight selection iterative wolf colony algorithm |
CN112365705A (en) * | 2020-08-27 | 2021-02-12 | 招商局重庆交通科研设计院有限公司 | Method for determining road traffic volume |
CN112463337A (en) * | 2020-12-08 | 2021-03-09 | 内蒙古大学 | Workflow task migration method used in mobile edge computing environment |
CN117373263A (en) * | 2023-12-08 | 2024-01-09 | 深圳市永达电子信息股份有限公司 | Traffic flow prediction method and device based on quantum pigeon swarm algorithm |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103646178A (en) * | 2013-12-18 | 2014-03-19 | 中国石油大学(华东) | Multi-objective optimization method based on improved gravitation search algorithm |
CN103678649A (en) * | 2013-12-20 | 2014-03-26 | 上海电机学院 | Traffic path searching system and method based on cloud self-adaptation particle swarm optimization |
CN103972908A (en) * | 2014-05-23 | 2014-08-06 | 国家电网公司 | Multi-target reactive power optimization method based on adaptive chaos particle swarm algorithm |
CN104008118A (en) * | 2013-04-23 | 2014-08-27 | 江南大学 | Method for improving population diversity in gravitational search algorithm |
CN104821082A (en) * | 2015-04-29 | 2015-08-05 | 电子科技大学 | Short-time traffic flow prediction method based on integrated evaluation |
CN105281847A (en) * | 2015-09-14 | 2016-01-27 | 杭州电子科技大学 | Deception jamming identification method based on model parameter identification |
CN105469611A (en) * | 2015-12-24 | 2016-04-06 | 大连理工大学 | Short-term traffic flow prediction model method |
CN105513350A (en) * | 2015-11-30 | 2016-04-20 | 华南理工大学 | Time-phased multi-parameter short-term traffic flow prediction method based on time-space characteristics |
CN106447024A (en) * | 2016-08-31 | 2017-02-22 | 上海电机学院 | Particle swarm improved algorithm based on chaotic backward learning |
-
2017
- 2017-06-26 CN CN201710495064.7A patent/CN107085942B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104008118A (en) * | 2013-04-23 | 2014-08-27 | 江南大学 | Method for improving population diversity in gravitational search algorithm |
CN103646178A (en) * | 2013-12-18 | 2014-03-19 | 中国石油大学(华东) | Multi-objective optimization method based on improved gravitation search algorithm |
CN103678649A (en) * | 2013-12-20 | 2014-03-26 | 上海电机学院 | Traffic path searching system and method based on cloud self-adaptation particle swarm optimization |
CN103972908A (en) * | 2014-05-23 | 2014-08-06 | 国家电网公司 | Multi-target reactive power optimization method based on adaptive chaos particle swarm algorithm |
CN104821082A (en) * | 2015-04-29 | 2015-08-05 | 电子科技大学 | Short-time traffic flow prediction method based on integrated evaluation |
CN105281847A (en) * | 2015-09-14 | 2016-01-27 | 杭州电子科技大学 | Deception jamming identification method based on model parameter identification |
CN105513350A (en) * | 2015-11-30 | 2016-04-20 | 华南理工大学 | Time-phased multi-parameter short-term traffic flow prediction method based on time-space characteristics |
CN105469611A (en) * | 2015-12-24 | 2016-04-06 | 大连理工大学 | Short-term traffic flow prediction model method |
CN106447024A (en) * | 2016-08-31 | 2017-02-22 | 上海电机学院 | Particle swarm improved algorithm based on chaotic backward learning |
Non-Patent Citations (2)
Title |
---|
李常洪等: "基于狼群算法优化的BP神经网络", 《科技创新与生产力》 * |
王健等: "BP与小波神经网络短时交通流预测对比研究", 《科技视界》 * |
Cited By (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107688702A (en) * | 2017-08-24 | 2018-02-13 | 西安理工大学 | A kind of river flood traffic camouflaging rule analogy method based on wolf pack algorithm |
CN107688702B (en) * | 2017-08-24 | 2020-11-17 | 西安理工大学 | Lane colony algorithm-based river channel flood flow evolution law simulation method |
CN107917734A (en) * | 2017-11-29 | 2018-04-17 | 国网吉林省电力有限公司信息通信公司 | Cable's Fault Forecasting Methodology based on temperature and resistance |
CN107917734B (en) * | 2017-11-29 | 2020-12-29 | 国网吉林省电力有限公司信息通信公司 | Optical cable fault prediction method based on temperature and resistance |
CN108491923A (en) * | 2018-04-10 | 2018-09-04 | 吉林大学 | Based on the pest image-recognizing method for improving wolf pack algorithm optimization Bayesian network |
CN108422997A (en) * | 2018-05-15 | 2018-08-21 | 南京航空航天大学 | A kind of automobile active safety cooperative control system and method based on wolf pack algorithm |
CN108918137A (en) * | 2018-06-08 | 2018-11-30 | 华北水利水电大学 | Fault Diagnosis of Gear Case devices and methods therefor based on improved WPA-BP neural network |
CN108898223A (en) * | 2018-07-11 | 2018-11-27 | 国家海洋技术中心 | A kind of ocean observation device operating status method for detecting abnormality and device |
CN109347900A (en) * | 2018-08-23 | 2019-02-15 | 东华理工大学 | Based on the cloud service system Adaptive evolution method for improving wolf pack algorithm |
CN109347900B (en) * | 2018-08-23 | 2020-12-11 | 东华理工大学 | Cloud service system self-adaptive evolution method based on improved wolf pack algorithm |
CN110444022A (en) * | 2019-08-15 | 2019-11-12 | 平安科技(深圳)有限公司 | The construction method and device of traffic flow data analysis model |
CN110824293A (en) * | 2019-10-15 | 2020-02-21 | 淮阴工学院 | Power grid fault diagnosis method based on multi-feature fusion parameters of wolf pack algorithm |
CN111259498A (en) * | 2020-01-14 | 2020-06-09 | 重庆大学 | Axle system thermal error modeling method and thermal error compensation system based on LSTM neural network |
CN111768622A (en) * | 2020-06-23 | 2020-10-13 | 南通大学 | Short-time traffic prediction method based on improved wolf algorithm |
CN111964574A (en) * | 2020-06-23 | 2020-11-20 | 安徽理工大学 | Three-dimensional laser target sphere center fitting method based on weight selection iterative wolf colony algorithm |
CN112365705A (en) * | 2020-08-27 | 2021-02-12 | 招商局重庆交通科研设计院有限公司 | Method for determining road traffic volume |
CN112365705B (en) * | 2020-08-27 | 2022-05-27 | 招商局重庆交通科研设计院有限公司 | Method for determining road traffic volume |
CN112463337A (en) * | 2020-12-08 | 2021-03-09 | 内蒙古大学 | Workflow task migration method used in mobile edge computing environment |
CN112463337B (en) * | 2020-12-08 | 2024-02-02 | 内蒙古大学 | Workflow task migration method used in mobile edge computing environment |
CN117373263A (en) * | 2023-12-08 | 2024-01-09 | 深圳市永达电子信息股份有限公司 | Traffic flow prediction method and device based on quantum pigeon swarm algorithm |
CN117373263B (en) * | 2023-12-08 | 2024-03-08 | 深圳市永达电子信息股份有限公司 | Traffic flow prediction method and device based on quantum pigeon swarm algorithm |
Also Published As
Publication number | Publication date |
---|---|
CN107085942B (en) | 2021-01-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107085942A (en) | A kind of traffic flow forecasting method based on wolf pack algorithm, apparatus and system | |
Chen et al. | Fuzzy forecasting based on fuzzy-trend logical relationship groups | |
CN107909206A (en) | A kind of PM2.5 Forecasting Methodologies based on deep structure Recognition with Recurrent Neural Network | |
CN106779148B (en) | A kind of method for forecasting wind speed of high speed railway line of multi-model multiple features fusion | |
CN110262233B (en) | Optimization method for technological parameters of magnetic control film plating instrument | |
CN108764540B (en) | Water supply network pressure prediction method based on parallel LSTM series DNN | |
CN103778482A (en) | Aquaculture dissolved oxygen short-term prediction method based on multi-scale analysis | |
CN107103397A (en) | A kind of traffic flow forecasting method based on bat algorithm, apparatus and system | |
CN106910337A (en) | A kind of traffic flow forecasting method based on glowworm swarm algorithm Yu RBF neural | |
Zeng | Neural computing in mechanics | |
CN107085941A (en) | A kind of traffic flow forecasting method, apparatus and system | |
CN107253195A (en) | A kind of carrying machine human arm manipulation ADAPTIVE MIXED study mapping intelligent control method and system | |
CN104317195B (en) | Improved extreme learning machine-based nonlinear inverse model control method | |
CN109800517B (en) | Improved reverse modeling method for magnetorheological damper | |
CN103927891A (en) | Crossroad dynamic turning proportion two-step prediction method based on double Bayes | |
CN105512832A (en) | Urban water demand combined predication method based on time-varying weight minimum variance | |
CN114357852A (en) | Layered water injection optimization method based on long-short term memory neural network and particle swarm optimization algorithm | |
CN108615097A (en) | A kind of wind speed forecasting method, system, equipment and computer readable storage medium | |
CN104504442A (en) | Neural network optimization method | |
CN105096007A (en) | Oil well yield prediction method and device based on improved neural network | |
CN109492768A (en) | Air Quality Forecast model training method and device | |
CN106991493A (en) | Sewage disposal water outlet parameter prediction method based on Grey production fuction | |
CN104680023B (en) | Oil pumper parameter optimization method based on multiobjective decision-making | |
El-Ghandour et al. | Groundwater management using a new coupled model of flow analytical solution and particle swarm optimization | |
CN111130909B (en) | Network flow prediction method based on self-adaptive reserve pool ESN |
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 |