CN108960486A - Interactive set evolvement method based on grey support vector regression prediction adaptive value - Google Patents
Interactive set evolvement method based on grey support vector regression prediction adaptive value Download PDFInfo
- Publication number
- CN108960486A CN108960486A CN201810603480.9A CN201810603480A CN108960486A CN 108960486 A CN108960486 A CN 108960486A CN 201810603480 A CN201810603480 A CN 201810603480A CN 108960486 A CN108960486 A CN 108960486A
- Authority
- CN
- China
- Prior art keywords
- individual
- adaptive
- population
- adaptive value
- support vector
- 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
- 230000003044 adaptive effect Effects 0.000 title claims abstract description 55
- 238000000034 method Methods 0.000 title claims abstract description 30
- 230000002452 interceptive effect Effects 0.000 title claims abstract description 24
- 230000035772 mutation Effects 0.000 claims abstract description 23
- 238000011156 evaluation Methods 0.000 claims abstract description 12
- 238000005259 measurement Methods 0.000 claims description 2
- 238000012706 support-vector machine Methods 0.000 claims description 2
- 241001282153 Scopelogadus mizolepis Species 0.000 claims 1
- 238000013461 design Methods 0.000 abstract description 6
- 238000005457 optimization Methods 0.000 description 18
- 230000008859 change Effects 0.000 description 8
- 230000014509 gene expression Effects 0.000 description 8
- 230000002068 genetic effect Effects 0.000 description 8
- 230000008901 benefit Effects 0.000 description 7
- 239000003795 chemical substances by application Substances 0.000 description 6
- 230000006399 behavior Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 241000894007 species Species 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 239000003086 colorant Substances 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000010429 evolutionary process Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- OYPRJOBELJOOCE-UHFFFAOYSA-N Calcium Chemical compound [Ca] OYPRJOBELJOOCE-UHFFFAOYSA-N 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000010353 genetic engineering Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000000116 mitigating effect Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 230000007474 system interaction Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
Classifications
-
- 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"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Biophysics (AREA)
- Business, Economics & Management (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Computational Linguistics (AREA)
- Game Theory and Decision Science (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Development Economics (AREA)
- Genetics & Genomics (AREA)
- General Health & Medical Sciences (AREA)
- Physiology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of interactive set evolutionary computation methods based on grey support vector regression prediction adaptive value, the specific steps of which are as follows: system provides design environment for user, generates initial Advanced group species at random before evolution starts;By interactive interface, people carries out numerical Evaluation to Nc individual;System clusters population according to individual similarity, and estimation clusters interior non-user and evaluates individual fitness, and predicts adaptive value using grey support vector regression;Set individual of evolving is carried out by Pareto is dominant sequence according to diversity, distributivity uncertainty measure, and set individual of evolving is operated using adaptive crossover and mutation, the generation interim population of same size;Parent population and interim population are merged, select top n individual as progeny population;Finally, being equidistantly divided into Nc unit to progeny population, 1 individual is randomly selected from each unit, Nc individual is recommended into user altogether.
Description
Technical field
The invention belongs to intelligence computation fields, predict adaptive value based on grey support vector regression more particularly to a kind of
Interactive mode set evolution optimization method, and it is used for the type selecting of color-match scheme.
Background technique
It is excellent that the Interactive evolutionary algorithm based on discovery learning that the 1990s proposes is applied to the implicit index of solution
Change problem needs to solve 2 basic problems: (1) how effectively to extract implicit knowledge;(2) how high quality, which solves implicit performance, refers to
Mark.
For the 1st problem, mainly there are 2 kinds of research strategies: first is that handing under the mode of evolution in interaction by people-machine
Mutual interface directly extracts implicit knowledge.This focuses primarily upon the research of adaptive value assignment mode, such as the periodical published in 2014
" automation journal " the 2nd phase " the weighting multi output Gaussian process agent model based on section adaptive value interactive genetic algorithm " adopts
With the uncertain number expression adaptive value such as interval number, reflect the preference characteristics of user;The periodical " electronic letters, vol " the published for 2017
12 phases " non-user based on maximum entropy formulism assigns adaptive value interactive genetic algorithm " and the periodical published in 2016
" Applied Intelligence " the 3rd phase " Predicting user ' s preferences using neural
Networks and psychology models " combines user browsing behavior to express individual demand, and it is suitable to make up value type
The deficiency of preference knowledge representation should be worth.Directly acquire that implicit knowledge calculation amount is small, and method is simple, but subjectivity is strong, and information is not true
Qualitative larger, noise content is higher.Second is that the implicit knowledge of indirect gain, i.e., excavate implicit knowledge during evolution, extraction has
It is worth preference information.Periodical " Zhengzhou University's journal (engineering version) " the 6th phase published for 2017, " was based on possible condition preference
The interactive genetic algorithm of network and its application " is fitted user preference using preference network, estimates that individual fitness and guidance are searched
Rope;Similarly, periodical " Automated Software Engineering " the 3rd phase " An published in 2017
Architecture based on interactive optimization and machine learning applied
To the next release problem " uses neural network learning user preference, estimates adaptive value.Indirect gain is implicit
Knowledge is excavated to the depth of preference information, enhances the search capability of algorithm, but agent model modeling comparison is complicated, estimation misses
Difference is unable to measure, and can still bring a large amount of adaptive value noises.Deeply excavate adaptive value estimation/expression strategy of user behavior pattern
Be expected to improve the effect or knowledge is extracted under more complicated scene that implicit knowledge is extracted, but integrally from optimization performance, only according to
There is significant limitations by improving implicit knowledge and extracting dynamics and obtain algorithm performance and promoted.For the 2nd problem, mainly pass through reduction
The efficient evolution strategy of human fatigue is realized.Algorithm search ability can be improved using big population scale, but this method needs
It solves not evaluating individual fitness estimation problem largely.Amount of user effort can be reduced by executing evolutional operation using agent model,
Time overhead is saved for user.The 8th phase of periodical " electronic letters, vol " published such as 2014 " is collected based on elite and selects individual of evolving
Interactive genetic algorithm " construct individual elite collection, similar individual classification is directly used in genetic manipulation with elite collection for selection, subtracts
Light burden for users.Above-mentioned thought is combined, algorithm performance can be significantly improved using evolution agent model under extensive population.
In above-mentioned algorithm, it is most important how to design Evolution of Population strategy, since agent model can generate error accumulation,
Controlling error is an insoluble problem.It evolves if set evolvement method is applied to extensive population cluster, it can be with
Improve search efficiency;Meanwhile having the estimation adaptive value of extensive population at individual as reappraised using agent model, then it can be with
Adaptive value precision is improved, the two fusion is learnt from other's strong points to offset one's weaknesses, then is expected to improve current interactive evolutionary optimization algorithm performance.Through consulting
Pertinent literature there is not yet the interactive set evolution optimization method using grey support vector regression prediction adaptive value at present.
As can developing relevant efficient design system, color matching design not only will push, will also have weight to other product designs
It is big to inspire meaning.
Summary of the invention
The technical problems to be solved by the present invention are: overcoming the shortcomings of existing adaptive value estimation technique, a kind of reduction is provided
Adaptive value evaluated error, the interactive mode for reducing burden for users, enhancing algorithm search ability and raising evolutionary optimization quality are evolved excellent
Change method.
The technical scheme is that
Firstly, expanding search space using extensive population, planted using limited user's evaluation individual as cluster centre
Clustering class.Then, to non-cluster center individual in clustering by the individual similarity measurement adaptive value for merging personal browsing behavior, and
Individual fitness is further predicted using grey support vector regression model, constitutes set evolution individual.Finally, using gather into
Change strategy and self-adaptive cross operation operation, set evolution algorithm is realized under NSGA-II normal form.To realize the present invention, it needs
It solves individual fitness prediction and set evolution strategy designs 2 main problems.
(1) grey support vector machines adaptive value prediction
By individual perfect number adaptive value f (xk(t)), xk(t) ∈ x (t) is denoted as F0=(f0(x1(t)),f0(x2(t)),…,f0
(xN(t)) individual fitness original series), are constituted.Then F01-AGO sequence be F1=(f1(x1(t)),f1(x2(t)),…,f1
(xN(t))),In order to make weighing factor one of the initial data in each characteristic dimension to objective function
It causes, original series is normalized first:
Data sequence is F after normalization0'=(f0'(x1(t)),f0'(x2(t)),…,f0'(xN(t))).Then, it establishes
F0' 1-AGO sequence F1=(f1'(x1(t)),f1'(x2(t)),…,f1'(xN(t))),Then
Grey supporting vector model is denoted as:
In formula: b is bias term;ω is weight.1-AGO converts item f1'(xk(t)) the individual fitness input space is represented
To the Nonlinear Mapping of high-dimensional feature space.The ε insensitive loss function that unknown parameter ω and b use Vapnik to propose in formula,
By the training set estimation in high-dimensional feature space:
Due to minimizingIt can guarantee ε minimum deflection, grey supporting vector model can be written as follow convex optimization problem:
In formula:The fitting precision of Controlling model;C is iotazation constant, controls the punishment to the sample beyond error
Degree.Above formula can be solved with Lagrange multiplier method:
Antithesis Optimized model can be obtained:
< f in formula1'(xi(t)),f1'(xj(t)) > representation vector inner product.The present invention selects kernel function K (f1'(xi
(t)),f1'(xj(t)))=exp (- υ | | f1'(xi(t)),f1'(xj(t)) | |), υ > 0, υ are nuclear parameters.Kernel function is substituted into
Above formula acquires αk,After b, grey support vector regression model are as follows:
In formula,It is the individual fitness f for newly inputting1'(xnew(t)) grey SVM prediction value.
Finally, predicted value is reverted to original series scale, obtain:
After population at individual cluster, adaptive value prediction, the individual subclass x of formation1(t),x2(t),…,xNcIt (t) is one by one
Individual subclass can be naturally enough considered as evolution individual (gathering individual of evolving) based on set evolution thought by body set.
(2) gather evolution strategy
By selecting suitable performance indicator, the optimization of implicit performance indicator can be converted to following general set decision and become
Measure optimization problem:
Max F (X)=(F1(X),F2(X),…,FId(X))
Wherein,For the power set of decision space S;X={ x1(t),x2(t),…,xNc(t) } kind constituted for individual of evolving
Group; Fd(X), d=1,2 ..., Id is the performance indicator of population X;Id is the dimension of optimization problem after conversion, and is much smaller than Nc.
In conjunction with implicit performance indicator feature, this section provides new set individual comparison measure, specific as follows.
Remember that t is for j-th of evolution individual of Advanced group species x (t)Wherein, xj(t) individual centered on, individual fitness isM=| xj(t) | for the individual x that evolvesj(t) number of individuals for including.
Using evolving, individual similarity information entropy portrays diversity:
In above formula,
Wherein, xir, r=1,2 ..., g are r attribute of composition individual,It is xirAttribute value.xj(t) center individual
xj(t) the cluster centre similarity μ (x individual with other evolutioni(t),xj(t)) bigger, xj(t) individual is evolved more with other
Similar, population diversity is poorer, at this point, F2(X) smaller.Conversely, F2(X) bigger, population is respectively evolved between individual more " loose ", is planted
Group's diversity is better.
Distributivity is portrayed using following formula:
In formula, d (xjIt (t)) is xj(t) minimum crowding distance,d*For kind
The average crowding distance of group's evolution individual,
For the evolution individual with identical sequence value, uncertainty is further portrayed using Interval Gray Number gray scale:
Above formula will evolution individual xj(t) adaptive value is considered as Interval Gray Number, xj(t) evaluation uncertainty can pass through area
Between degree of greyness of grey number portray, i.e. GF (X) is smaller, evolve individual uncertainty it is smaller, vice versa.
Work as x1||sparx2When, select x=arg min { GF (x1),GF(x2) it is used as winning individual.Pass through the above method, energy
Enough any 2 set evolution individuals to population carry out superiority and inferiority comparison.
(3) adaptive set intersection and mutation probability
For set evolution feature, this trifle provides adaptive set and intersects and mutation operation operator.For same evolution
Crossover probability inside individual, using adaptive crossover mutation:
Above formula considers in evolutionary process between crossover operation and set evolution diversity of individuals and uncertain variation
Relationship.In order to enhance search capability, the uncertainty for individual of evolving should be directly proportional to crossover probability;In order to retain advantage
The diversity of body, individual of evolving should be inversely proportional with crossover probability, and Integral cross probability should be gradually reduced with evolution, reinforce algorithm and receive
Holding back property;Vice versa.
Set evolves individual mutation operation using single-point variation mode, and self-adaptive mutation is by Pm 1And Pm 22 part structures
At:
Above formula embodies the diversity and constringent relationship of mutation operation and individual of evolving.In order to retain advantage
Body, diversity should be inversely proportional with convergence with mutation probability, vice versa.
In formulaIt is the individual x that evolvesj(t) to variation individual inAdaptive value.Above formula embody mutation operation with
The relationship of evolution individual fitness.It as protects advantage individual not to be destroyed, selects the evolution individual x of mutation operationj(t) in,
The bigger individual of adaptive value is performed the probability of variation with regard to smaller.
Evolution Individual Adaptive mutation probability PmAre as follows:
Color individual UVR exposure method is: RGB color mode passes through the change to red (R), green (G), blue (B) three color attributes
Change and their superposition obtains various colors, each color attribute value range is 0~255.RGB color individual dye
Colour solid uses binary coding, and code length is 24, wherein the red attribute of preceding 8 expressions, intermediate 8 expressions green belong to
Property, last 8 expressions blueness attribute, it is 00000000~11111111 that each color attribute, which corresponds to binary coding range,.It is excellent
Changing target is certain color of object being previously set, and by evolutionary optimization, user is obtained to a with the matched color of color of object
Body.
The advantages and positive effects of the present invention are:
1, the present invention evaluates individual fitness using grey SVM prediction non-user, estimates plan compared to other adaptive values
Slightly, evaluated error is greatly reduced, while alleviating human fatigue;
2, the present invention is realized using new set evolution strategy and is evolved, hence it is evident that is improved search efficiency and optimization quality, is answered
Good effect of optimization is obtained for color-match problem.
Detailed description of the invention
Fig. 1 is overview flow chart of the invention;
Fig. 2 is system interaction surface chart of the invention;
Fig. 3 is search time contrast schematic diagram of the invention;
Fig. 4 is evolutionary generation contrast schematic diagram of the invention.
Specific embodiment
- Fig. 4 referring to Fig.1 does and is described in detail further below to present invention implementation, following embodiment be it is descriptive, be not
Limited, this does not limit the scope of protection of the present invention.
It is a kind of to predict the interactive set evolutionary computation process of adaptive value as shown in Fig. 1 based on grey support vector regression.
The step of this method, is as follows:
Step 1. interactive interface is made of 3 parts, and part 1 is the evolution module positioned at interface left side, which is in
12 individuals (color block) of user are now given, each individual lower section setting adaptive value inputs text box.Meanwhile system passes through under individual
Square sliding shoe records user to the evaluation time of each individual, for calculating adaptive value uncertainty.It is a external for convenient for comparing
Setting color of object block is enclosed, and shows the distance value with current individual.Part 2 be positioned at interface upper right quarter evolution parameter and
Information Statistics module, user can input RGB color value by dragging scroll bar before evolving, and color of object is arranged.In addition, the mould
Block also shows the statistical informations such as start time of evolving, current evolutionary generation, user's evaluation number of individuals, evaluation total time-consuming.3rd
Dividing is the order button module for being located at interface right lower quadrant, user click START button, and system initialization simultaneously generates initial population;
After user's evaluation individual, click the Next button, system clustered from the background, adaptive value estimation, set are evolved etc. evolves
Operation generates next-generation Advanced group species.The process is recycled, until meeting termination condition, " end " button is clicked and terminates evolution.
Color individual UVR exposure method is: RGB color mode passes through the change to red (R), green (G), blue (B) three color attributes
Change and their superposition obtains various colors, each color attribute value range is 0~255.RGB color individual dye
Colour solid uses binary coding, and code length is 24, wherein the red attribute of preceding 8 expressions, intermediate 8 expressions green belong to
Property, last 8 expressions blueness attribute, it is 00000000~11111111 that each color attribute, which corresponds to binary coding range,.It is excellent
Changing target is certain color of object being previously set, and by evolutionary optimization, user is obtained to a with the matched color of color of object
Body.
Step 2.
Firstly, selected section individual xi(t), i=1,2 ... Nc are individual as cluster centre, and by user's evaluation adaptive value
f(xi(t)).Then, the non-central individual x of other in population is calculated one by one by formulao(t) with the similarity of each center individual, by similar
Degree clusters individual, is denoted as xi(t)={ xi(t)},i∈{1,2,…Nc}.In this way, population x (t) is finally divided into Nc subclass,
It is denoted as x1(t),x2(t),…,xNc(t).Then, the weighted average with each cluster centre similarity is sought in each individual subclass
Value estimates non-central individual xo(t) adaptive value.In order to reduce adaptive value evaluated error, to non-central individual estimation adaptive value f (xo
(t)) using grey support vector regression prediction, the non-central final adaptive value of individual for subsequent evolution is obtained
Step 3.
Diversity, distributivity and uncertainty measure are successively calculated to set evolution individual.
Diversity measure:
Distributivity is estimated:
In formula, d (xjIt (t)) is xj(t) minimum crowding distance,d*For population
The average crowding distance of individual of evolving,
Uncertainty measure:
Above formula will evolution individual xj(t) adaptive value is considered as Interval Gray Number, xj(t) evaluation uncertainty can pass through area
Between degree of greyness of grey number portray, i.e. GF (X) is smaller, evolve individual uncertainty it is smaller, vice versa.
Work as x1||sparx2When, select x=arg min { GF (x1),GF(x2) it is used as winning individual.Pass through the above method, energy
Enough any 2 evolution individuals to population carry out superiority and inferiority comparison.
Step 4.
Individual is evolved using adaptive crossover and mutation operation to set, generates the interim population of same size.It is adaptive to hand over
Pitch probability:
Above formula considers in evolutionary process between crossover operation and set evolution diversity of individuals and uncertain variation
Relationship.In order to enhance search capability, the uncertainty for individual of evolving should be directly proportional to crossover probability;In order to retain advantage
The diversity of body, individual of evolving should be inversely proportional with crossover probability, and Integral cross probability should be gradually reduced with evolution, reinforce algorithm and receive
Holding back property;Vice versa.
Set evolve individual mutation operation using single-point make a variation mode, self-adaptive mutation byWithPart structure
At:
Above formula embodies mutation operation and the evolve diversity of individual and the relationship of distributivity.In order to retain advantage
Body, diversity should be inversely proportional with distributivity with mutation probability, vice versa.
In formulaIt is the individual x that evolvesj(t) to variation individual inAdaptive value.To protect advantage individual not broken
It is bad, select the evolution individual x of mutation operationj(t) in, the bigger individual of adaptive value is performed the probability of variation with regard to smaller.
Evolution Individual Adaptive mutation probability PmAre as follows:
Step 5.
Parent population and interim population are merged, and to population sequence is merged, select top n individual as progeny population;
Finally, being equidistantly divided into Nc unit to progeny population, 1 individual is randomly selected from each unit, altogether by Nc individual
Recommend user.
The present invention is compared with current algorithm
By " interactive genetic algorithm based on possible condition preference network " (Probabilistic Conditional
Preference Network Assisted Interactive Genetic Algorithm, PCPN-IGA) " and " based on essence
The interactive genetic algorithm of English collection selection evolution individual " (Interactive Genetic Algorithms with
Selecting Individuals Using Elite Set, IGA-SES) etc. two kinds of related algorithms as comparison algorithm, verifying
Validity of present invention in terms of search efficiency, optimization quality, mitigation.
The experimental result of three kinds of methods is as shown in Figure 4.The present invention is substantially better than comparison algorithm as can be seen from Figure 4, specifically
It shows themselves in that
A. in terms of search time, the time that the present invention expends is minimum, improves search efficiency.Reason is that the present invention adopts
With set evolution strategy and NSGA-II algorithm engine, Species structure can be made more uniform, diversity is more preferable, improves algorithm
Search efficiency.Convergence is more strengthened using adaptive crossover and mutation operation.Although for gathering individual ratio of evolving
Compared with expending, the time is longer, but whole search time is still shorter than comparison algorithm.
B. in terms of evolutionary generation, evolutionary generation of the present invention is minimum, and which not only improves search efficiencies, also reduces user
Fatigue.Reason is that the present invention estimates individual fitness using the individual similarity based on user browsing behavior, while using ash
Support vector regression further predicts adaptive value, improves adaptive value precision, so Evolutionary direction more meets the preference of people,
Accelerate algorithmic statement.
Following table gives the search result mean value and exact matching number of users statistics of three kinds of methods.Data in table are shown
The Mann-Whitney U that work property level is 0.05 is examined.As can be seen from the table, refer in evaluation number of individuals, Search of Individual number two
Put on, minimum number of the invention, and with control methods significant difference.In optimal skill, although quantity of the invention with it is right
Ratio method difference is not significant, but still is most.This shows that the present invention (including can evolve in relatively minimal evaluation quantity
Algebra) under, obtain most satisfactory solutions.This reflects optimization efficiency highest of the invention.Search of Individual quantity of the invention compared with
It is caused by evolutionary generation is less less.Because control methods is also the extensive Evolution of Population of identical population scale, evolutionary generation is got over
More, Search of Individual number is also more, so, the Search of Individual number of control methods is more cannot to illustrate search capability better than the present invention.
It is exactly matched on solution number of users obtaining, the present invention has 3, and IGA-SES has 2, and PCPN-IGA only has 1.This explanation,
This complicated implicit performance indicator optimization problem for color-match, what most users finally obtained is still approximate solution.But this
The number of users that invention can obtain exact matching solution is most, reflects that optimization performance of the invention is best.It, can be with by above-mentioned analysis
See that search performance of the invention is best.
A. the present invention > IGA-SES > PCPN-IGA b. the present invention=IGA-SES=PCPN-IGA
The above described is only a preferred embodiment of the present invention, be not intended to limit the present invention in any form, it is all
It is any simple modification, equivalent change and modification to the above embodiments according to the technical essence of the invention, still falls within
In the range of technical solution of the present invention.
Claims (4)
1. a kind of interactive set evolvement method based on grey support vector regression prediction adaptive value, it is characterized in that: using big
Scale population expands search space, carries out population cluster by cluster centre of limited user's evaluation individual;Then, in cluster
Non-cluster center individual uses grey support vector regression by the individual similarity measurement adaptive value for merging individual's browsing behavior
Model further predicts individual fitness, constitutes set evolution individual;Finally, being become using set evolution strategy and adaptive intersection
ETTHER-OR operation realizes that set is evolved;
(1) grey support vector machines adaptive value prediction
For the individual fitness f newly inputted1'(xnew(t)) grey SVM prediction value are as follows:
(2) gather evolution strategy
Diversity measure:
Distributivity is estimated:
Uncertainty measure:
(3) adaptive set intersection and mutation probability
Adaptive crossover mutation:
Self-adaptive mutation:
2. the interactive set evolvement method according to claim 1 based on grey support vector regression prediction adaptive value,
It is characterized in that:
3. the interactive set evolutionary computation side according to claim 1 based on grey support vector regression prediction adaptive value
Method, it is characterized in that: d (xjIt (t)) is xj(t) minimum crowding distance,d*For population
The average crowding distance of individual of evolving,
4. the interactive set evolvement method according to claim 1 based on grey support vector regression prediction adaptive value,
It is characterized in that: self-adaptive mutation PmByWith2 parts are constituted,
In formulaIt is the individual x that evolvesj(t) to variation individual inAdaptive value.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810603480.9A CN108960486B (en) | 2018-06-12 | 2018-06-12 | Interactive set evolution method for predicting adaptive value based on gray support vector regression |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810603480.9A CN108960486B (en) | 2018-06-12 | 2018-06-12 | Interactive set evolution method for predicting adaptive value based on gray support vector regression |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108960486A true CN108960486A (en) | 2018-12-07 |
CN108960486B CN108960486B (en) | 2021-10-29 |
Family
ID=64488560
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810603480.9A Expired - Fee Related CN108960486B (en) | 2018-06-12 | 2018-06-12 | Interactive set evolution method for predicting adaptive value based on gray support vector regression |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108960486B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110162704A (en) * | 2019-05-21 | 2019-08-23 | 西安电子科技大学 | More scale key user extracting methods based on multiple-factor inheritance algorithm |
CN110276375A (en) * | 2019-05-14 | 2019-09-24 | 嘉兴职业技术学院 | A kind of identification and processing method of crowd's dynamic clustering information |
CN110739045A (en) * | 2019-10-14 | 2020-01-31 | 郑州航空工业管理学院 | Interactive evolution optimization method for personalized recipe design |
CN115204064A (en) * | 2022-09-19 | 2022-10-18 | 中国石油大学(华东) | Gas injection huff-puff parameter optimization method and system in shale oil exploitation |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102270302A (en) * | 2011-07-20 | 2011-12-07 | 北京航空航天大学 | Grey support vector machine-based multi-stress accelerated life testing forecasting method |
US20140245207A1 (en) * | 2013-02-25 | 2014-08-28 | Christian D. Poulin | Interfaces for predictive models |
CN104596767A (en) * | 2015-01-13 | 2015-05-06 | 北京工业大学 | Method for diagnosing and predicating rolling bearing based on grey support vector machine |
CN104749219A (en) * | 2013-07-30 | 2015-07-01 | 中国标准化研究院 | A honey detection method by support vector machine classifier parameter selection based on particle swarm optimization |
CN106296444A (en) * | 2016-07-27 | 2017-01-04 | 国网山东省电力公司经济技术研究院 | A kind of annual whole society based on particle cluster algorithm and support vector machine power predicating method |
CN106845136A (en) * | 2017-02-21 | 2017-06-13 | 上海大学 | A kind of needle-valve body crush and grind accuracy prediction method based on SVMs |
CN107609681A (en) * | 2017-08-22 | 2018-01-19 | 西安建筑科技大学 | A kind of more metal multiple target ore-proportioning methods based on APSO algorithm |
US20180082192A1 (en) * | 2016-09-21 | 2018-03-22 | Scianta Analytics, LLC | Cognitive modeling apparatus including multiple knowledge node and supervisory node devices |
CN107993723A (en) * | 2017-12-19 | 2018-05-04 | 苏州大学 | A kind of warfarin dose prediction modeling method based on integrated evolutionary learning |
CN108022014A (en) * | 2017-12-05 | 2018-05-11 | 国网山东省电力公司济南供电公司 | A kind of Load Prediction In Power Systems method and system |
-
2018
- 2018-06-12 CN CN201810603480.9A patent/CN108960486B/en not_active Expired - Fee Related
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102270302A (en) * | 2011-07-20 | 2011-12-07 | 北京航空航天大学 | Grey support vector machine-based multi-stress accelerated life testing forecasting method |
US20140245207A1 (en) * | 2013-02-25 | 2014-08-28 | Christian D. Poulin | Interfaces for predictive models |
CN104749219A (en) * | 2013-07-30 | 2015-07-01 | 中国标准化研究院 | A honey detection method by support vector machine classifier parameter selection based on particle swarm optimization |
CN104596767A (en) * | 2015-01-13 | 2015-05-06 | 北京工业大学 | Method for diagnosing and predicating rolling bearing based on grey support vector machine |
CN106296444A (en) * | 2016-07-27 | 2017-01-04 | 国网山东省电力公司经济技术研究院 | A kind of annual whole society based on particle cluster algorithm and support vector machine power predicating method |
US20180082192A1 (en) * | 2016-09-21 | 2018-03-22 | Scianta Analytics, LLC | Cognitive modeling apparatus including multiple knowledge node and supervisory node devices |
CN106845136A (en) * | 2017-02-21 | 2017-06-13 | 上海大学 | A kind of needle-valve body crush and grind accuracy prediction method based on SVMs |
CN107609681A (en) * | 2017-08-22 | 2018-01-19 | 西安建筑科技大学 | A kind of more metal multiple target ore-proportioning methods based on APSO algorithm |
CN108022014A (en) * | 2017-12-05 | 2018-05-11 | 国网山东省电力公司济南供电公司 | A kind of Load Prediction In Power Systems method and system |
CN107993723A (en) * | 2017-12-19 | 2018-05-04 | 苏州大学 | A kind of warfarin dose prediction modeling method based on integrated evolutionary learning |
Non-Patent Citations (1)
Title |
---|
赵玲 等: ""基于过程支持向量回归机的油田开发指标预测模型"", 《数学的实践与认识》 * |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110276375A (en) * | 2019-05-14 | 2019-09-24 | 嘉兴职业技术学院 | A kind of identification and processing method of crowd's dynamic clustering information |
CN110276375B (en) * | 2019-05-14 | 2021-08-20 | 嘉兴职业技术学院 | Method for identifying and processing crowd dynamic clustering information |
CN110162704A (en) * | 2019-05-21 | 2019-08-23 | 西安电子科技大学 | More scale key user extracting methods based on multiple-factor inheritance algorithm |
CN110162704B (en) * | 2019-05-21 | 2022-06-10 | 西安电子科技大学 | Multi-scale key user extraction method based on multi-factor genetic algorithm |
CN110739045A (en) * | 2019-10-14 | 2020-01-31 | 郑州航空工业管理学院 | Interactive evolution optimization method for personalized recipe design |
CN110739045B (en) * | 2019-10-14 | 2023-05-30 | 郑州航空工业管理学院 | Interactive evolution optimization method for personalized recipe design |
CN115204064A (en) * | 2022-09-19 | 2022-10-18 | 中国石油大学(华东) | Gas injection huff-puff parameter optimization method and system in shale oil exploitation |
Also Published As
Publication number | Publication date |
---|---|
CN108960486B (en) | 2021-10-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Trivedi et al. | A survey of multiobjective evolutionary algorithms based on decomposition | |
CN108960486A (en) | Interactive set evolvement method based on grey support vector regression prediction adaptive value | |
Purshouse | On the evolutionary optimisation of many objectives | |
CN107465664B (en) | Intrusion detection method based on parallel multi-worker bee colony algorithm and support vector machine | |
CN109461025A (en) | A kind of electric energy substitution potential customers' prediction technique based on machine learning | |
CN110990718B (en) | Social network model building module of company image lifting system | |
CN111061959B (en) | Group intelligent software task recommendation method based on developer characteristics | |
Ding et al. | Product color emotional design considering color layout | |
CN109492748B (en) | Method for establishing medium-and-long-term load prediction model of power system based on convolutional neural network | |
CN111343650B (en) | Urban scale wireless service flow prediction method based on cross-domain data and loss resistance | |
CN105184415B (en) | A kind of Distribution Networks Reconfiguration design method | |
CN109543939A (en) | A kind of method of green building productions certification risk evaluation model building | |
Gong et al. | Set-based many-objective optimization guided by a preferred region | |
CN116526450A (en) | Error compensation-based two-stage short-term power load combination prediction method | |
CN115912502A (en) | Intelligent power grid operation optimization method and device | |
CN108737429B (en) | Network intrusion detection method | |
Li et al. | A hybrid coevolutionary algorithm for designing fuzzy classifiers | |
Gupta et al. | Image enhancement using ant colony optimization | |
Zhang et al. | Embedding multi-attribute decision making into evolutionary optimization to solve the many-objective combinatorial optimization problems | |
CN113688960A (en) | Grey wolf optimization GHFCM-based residential power data clustering method and device | |
CN113987808A (en) | Electricity user complaint early warning method of feature weighted Bayesian network | |
CN117313795A (en) | Intelligent building energy consumption prediction method based on improved DBO-LSTM | |
CN100380395C (en) | Supervised classification process of artificial immunity in remote sensing images | |
CN110705756A (en) | Electric power energy consumption optimization control method based on input convex neural network | |
CN110059752A (en) | A kind of statistical learning querying method based on comentropy Sampling Estimation |
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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20211029 |