CN110033053A - Optimize the Favorable Reservoir prediction technique of twin support vector machines based on improvement drosophila - Google Patents
Optimize the Favorable Reservoir prediction technique of twin support vector machines based on improvement drosophila Download PDFInfo
- Publication number
- CN110033053A CN110033053A CN201910323887.0A CN201910323887A CN110033053A CN 110033053 A CN110033053 A CN 110033053A CN 201910323887 A CN201910323887 A CN 201910323887A CN 110033053 A CN110033053 A CN 110033053A
- Authority
- CN
- China
- Prior art keywords
- drosophila
- favorable
- favorable reservoir
- prediction model
- reservoir
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000002349 favourable effect Effects 0.000 title claims abstract description 53
- 241000255581 Drosophila <fruit fly, genus> Species 0.000 title claims abstract description 46
- 238000000034 method Methods 0.000 title claims abstract description 18
- 238000012706 support-vector machine Methods 0.000 title claims abstract description 13
- 230000006872 improvement Effects 0.000 title claims abstract description 8
- 238000005457 optimization Methods 0.000 claims abstract description 10
- 238000012549 training Methods 0.000 claims abstract description 6
- 230000008569 process Effects 0.000 abstract description 4
- 238000011161 development Methods 0.000 description 18
- 239000000796 flavoring agent Substances 0.000 description 9
- 235000019634 flavors Nutrition 0.000 description 9
- 238000010801 machine learning Methods 0.000 description 4
- 241000208340 Araliaceae Species 0.000 description 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 2
- 235000003140 Panax quinquefolius Nutrition 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 235000008434 ginseng Nutrition 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007634 remodeling Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06N3/004—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]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of Favorable Reservoir prediction techniques for optimizing twin support vector machines based on improvement drosophila, it is characterised in that by matching seismic attributes data and Reservoir levels label, obtains the earthquake sample set with category label；Twin support vector machine classifier is constructed between any two class of Favorable Reservoir, reservoir and non-Favorable Reservoir according to " one-versus-one " strategy, and then group is combined into Favorable Reservoir prediction model；Determining problem is difficult to for parameter in prediction model training process, the punishment parameter optimized and nuclear parameter will be needed to be set as drosophila coordinate value first, random initializtion drosophila group, secondly drosophila group is divided by multiple subgroups according to certain rule, each subgroup is evolved respectively, can expand the search range of drosophila, using the classification accuracy of Favorable Reservoir prediction model as foundation, superiority and inferiority judgement, optimization update are carried out to parameter, finally obtain the optimized parameter of Favorable Reservoir prediction model；Using the seismic properties sample of zone of ignorance as input, favo(u)rable target is drawn a circle to approve using the Favorable Reservoir prediction model after optimization.The present invention is determined optimal Favorable Reservoir prediction model, can effectively improve the accuracy of reservoir prediction by being optimized based on the improvement drosophila algorithm shared out the work and help one another.
Description
Technical field
The invention belongs to field of geophysical exploration and artificial intelligence field, and in particular to one kind is optimized based on drosophila is improved
The Favorable Reservoir prediction technique of twin support vector machines.
Background technique
With the development of exploration engineering, the ratio of oil-gas reservoir has huge Exploration Potential in continuous enlargement.Favorable Reservoir
Prediction is to provide the process of geologic basis for well site deployment and development plan, and effectively prediction can save a large amount of manpower, object
Power and financial resource.Therefore, Favorable Reservoir prediction is the key link during exploration and development.Traditional prediction technique generally according to
Rely artificial experience, the seismic properties utilized are few, cause precision of prediction not high.With the development of computer technology, in machine learning
Related algorithm in relation to classification prediction is that geological personnel provides new approaches in terms of Favorable Reservoir prediction, using machine learning side
Method prediction Favorable Reservoir can excavate the mapping relations between seismic properties and Favorable Reservoir, so that being applied to other does not explore area
Domain, auxiliary seismic prospecting personnel quickly draw a circle to approve Favorable Areas.
But since the accuracy rate of most of classification prediction algorithms is influenced by parameter selection, determine the optimal ginseng of prediction model
Number can improve the precision of Favorable Reservoir prediction to a certain extent.Therefore, a kind of Favorable Reservoir that parameter is optimal how is designed
Prediction technique becomes problem urgently to be resolved in exploration process.
Summary of the invention
In order to realize that Favorable Reservoir can be effectively predicted in machine learning method, while solving parameter during predicting and being difficult to really
Fixed problem, the invention proposes the Favorable Reservoir prediction techniques for optimizing twin support vector machines based on improvement drosophila, by adopting
Classified twin support vector machines with improved drosophila algorithm optimization more, constructs advantageous reservoir prediction model, assist seismic prospecting people
Member quickly draws a circle to approve Favorable Areas.
To achieve the above object, technical solution of the present invention mainly includes the following steps:
A. data prediction:
Seismic properties and lithological profile data are extracted from the data sources such as exploration database, seismic data cube, pass through well
Data Matching is shaken, the earthquake sample set for having class label is obtained.Wherein, reservoir type is divided into Favorable Reservoir development area, reservoir
Development area and non-Favorable Reservoir development area.
B. advantageous reservoir prediction model is constructed:
Using " one-versus-one " as prediction model construction strategy, first in Favorable Reservoir development area, reservoir development
Area, non-Favorable Reservoir development area any two class between the two twin support vector machines of classification of training, pass through the 3* (3- of combination producing
1)/2=3 two twin the support vector machines of classification constructs advantageous reservoir prediction model.
C. prediction model optimized parameter determines:
(1) initialization model parameter
The punishment parameter of Favorable Reservoir prediction model and nuclear parameter are set as drosophila coordinate value by the present invention, and drosophila is arranged
The maximum number of iterations and population scale size of algorithm assign initial position, each position component generation respectively for each drosophila individual
Two punishment parameters and nuclear parameter of Favorable Reservoir prediction model when table is initial, and assign each drosophila an individual distance at random
The direction and.
(2) model parameter is updated
Since the value of optimized parameter is unknown, by estimating position of each drosophila apart from origin, flavor concentration judgement is calculated
Value calculates corresponding flavor concentration according to the classification accuracy of Favorable Reservoir prediction model.
Under the inspiration that today's society group shares out the work and help one another, drosophila group is divided into multiple sons according to certain rules
Group, each subgroup evolves respectively, can expand the search range of drosophila, be easy to jump out local optimum.The present invention by drosophila group with
Machine is divided into 3 subgroups, finds the corresponding drosophila position of maximum flavor concentration in each subgroup, records the optimal taste in each subgroup
Then the position of drosophila individual in each subgroup is updated to this 3 optimal locations by concentration value and corresponding position respectively.
(3) model optimized parameter is determined
It is randomly provided a distance and direction for the drosophila in each subgroup again, judges whether to reach maximum number of iterations,
If not up to, continuing to execute step (2), otherwise, the corresponding parameter of global best flavors concentration value being determined as Favorable Reservoir
The optimized parameter of prediction model.
D. it is predicted using the Favorable Reservoir prediction model after optimization:
Using the seismic properties sample set with class label as training sample, mould is predicted using the Favorable Reservoir after optimization
Type is trained, and is predicted using Favorable Reservoir development area of the trained model to zone of ignorance.
The beneficial effects of the present invention are: predictablity rate can be improved using machine learning method prediction Favorable Reservoir, but
Most of prediction models are largely influenced by parameter selection, in order to advanced optimize Favorable Reservoir prediction model, this
Invention is based on reservoir specific diversity, constructs more twin support vector cassification models of classification, and close based on the division of labor using a kind of
The improvement drosophila algorithm optimization model parameter for making thought greatly improves the validity of Favorable Reservoir prediction, is that geological personnel is quick
Delineation Favorable Areas provides booster action.
Detailed description of the invention
Fig. 1 is flow chart of the invention
Specific embodiment
Below with reference to Fig. 1, the present invention is described in further detail:
A. data prediction:
Seismic properties and lithological profile data are extracted from the data sources such as exploration database, seismic data cube, using rule
Generalized method pre-processes seismic properties, shakes Data Matching by well, obtains the earthquake sample set for having class label.Its
In, the number of seismic properties is m, and reservoir type is divided into Favorable Reservoir development area, reservoir development area and non-Favorable Reservoir development area,
It is denoted as 0,1,2 respectively.The present invention uses ten folding cross validations, and data set is divided into ten parts, wherein nine parts are used to train, it is a
For testing.
B. advantageous reservoir prediction model is constructed:
Using " one-versus-one " as prediction model construction strategy, first in Favorable Reservoir development area, reservoir development
Area, non-Favorable Reservoir development area any two class between the two twin support vector machines of classification of training, pass through the 3* (3- of combination producing
1)/2=3 two twin the support vector machines of classification constructs advantageous reservoir prediction model.
C. prediction model optimized parameter determines:
(1) initialization model parameter
The punishment parameter of Favorable Reservoir prediction model and nuclear parameter are set as drosophila coordinate value by the present invention, and drosophila is calculated
The maximum number of iterations of method is set as iter_max, and population scale is dimensioned to N, assigns initial position for each drosophila individual
(Cinit1,Cinit2,δinit), wherein Cinit1,Cinit2,δinitTwo punishment of Favorable Reservoir prediction model when respectively representing initial
Parameter and nuclear parameter, and assign the individual distance of each drosophila and direction at random according to following formula.
Ci1=Cinit1+RandomValue
Ci2=Cinit2+RandomValue
δi=δinit+RandomValue
I=1,2 ..., N
Wherein Ci1,Ci2,δiTwo punishment parameters and nuclear parameter representated by respectively drosophila individual i, i indicate drosophila individual
Number.
(2) model parameter is updated
Since the value of optimized parameter is unknown, the present invention is by estimating position (Dist of each drosophila apart from origini), it calculates
Flavor concentration judgment value (Si), this value is the inverse of distance.According to the classification accuracy of Favorable Reservoir prediction model
Function calculates corresponding flavor concentration (Smelli)。
Si=1/Disti
Smelli=Function (Si)
Although drosophila algorithm is simple, efficient, only learn in an iterative process to optimal drosophila individual, if the drosophila generation
The position of table is not global optimum, then is easily trapped into local optimum, influences convergence precision and speed.For drosophila algorithm
Drosophila group is divided into multiple by deficiency, the present invention according to certain rules under the inspiration that today's society group shares out the work and help one another
Subgroup, each subgroup are evolved respectively, can expand the search range of drosophila, are easy to jump out local optimum.
Drosophila group random division is 3 subgroups by the present invention, finds the corresponding drosophila of maximum flavor concentration in each subgroup
Position is denoted as bestIndex1,bestIndex2,bestIndex3.Record the optimal flavor concentration value in each subgroup
SmellbestjAnd corresponding position Ci1(bestIndexj),Ci2(bestIndexj),δi(bestIndexj), then by each son
The position of drosophila individual is updated to this 3 optimal locations respectively in group.
Smellbestj=bestIndexj
Cj1=Ci1(bestIndexj)
Cj2=Ci2(bestIndexj)
δj=δi(bestIndexj)
J=1,2,3
Wherein Cj1,Cj2,δjTwo punishment parameters and the core ginseng that drosophila individual represents in respectively updated j-th of subgroup
Number, j indicate the number of drosophila subgroup.
(3) model optimized parameter is determined
It is randomly provided a distance and direction for the drosophila in each subgroup again, judges whether to reach maximum number of iterations
Iter_max, if not up to, step (2) are continued to execute, otherwise, by global best flavors concentration value-max (Smellbest1,
Smellbest2,Smellbest3) corresponding parameter is determined as the optimized parameter of Favorable Reservoir prediction model.
D. it is predicted using the Favorable Reservoir prediction model after optimization:
Using the seismic properties sample set with class label as training sample, mould is predicted using the Favorable Reservoir after optimization
Type is trained, and is predicted using Favorable Reservoir development area of the trained model to zone of ignorance.
The above is only presently preferred embodiments of the present invention, and any person skilled in the art is possibly also with above-mentioned
The equivalent example of equivalent variations is retrofited or be changed to the technical solution of elaboration.It is all without departing from technical solution of the present invention content,
Any simple modification, change or the remodeling that technical solution according to invention carries out above-described embodiment, belong to inventive technique side
The protection scope of case.
Claims (1)
1. optimizing the Favorable Reservoir prediction technique of twin support vector machines based on improvement drosophila, which is characterized in that including following step
It is rapid:
Seismic properties and lithological profile data are extracted from the data sources such as exploration database, seismic data cube, and number is shaken by well
According to matching, the earthquake sample set for having class label is obtained；According to " one-versus-one " strategy Favorable Reservoir, reservoir and
Twin support vector machine classifier is established between any two class of non-Favorable Reservoir, constructs advantageous reservoir prediction model；For pre-
Parameter is difficult to determining problem during surveying model training, is carried out using a kind of improvement drosophila algorithm based on thought of sharing out the work and help one another
Drosophila group is divided into multiple subgroups according to certain rule by model optimization, this method, and each subgroup is evolved respectively, can be expanded
The search range of drosophila obtains the optimized parameter of Favorable Reservoir prediction model；Using the seismic properties sample of zone of ignorance as defeated
Enter, favo(u)rable target is drawn a circle to approve using the Favorable Reservoir prediction model after optimization.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910323887.0A CN110033053A (en) | 2019-04-22 | 2019-04-22 | Optimize the Favorable Reservoir prediction technique of twin support vector machines based on improvement drosophila |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910323887.0A CN110033053A (en) | 2019-04-22 | 2019-04-22 | Optimize the Favorable Reservoir prediction technique of twin support vector machines based on improvement drosophila |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110033053A true CN110033053A (en) | 2019-07-19 |
Family
ID=67239734
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910323887.0A Pending CN110033053A (en) | 2019-04-22 | 2019-04-22 | Optimize the Favorable Reservoir prediction technique of twin support vector machines based on improvement drosophila |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110033053A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105718932A (en) * | 2016-01-20 | 2016-06-29 | 中国矿业大学 | Colorful image classification method based on fruit fly optimization algorithm and smooth twinborn support vector machine and system thereof |
| CN109002918A (en) * | 2018-07-16 | 2018-12-14 | 国网浙江省电力有限公司经济技术研究院 | Based on drosophila optimization algorithm-support vector machines electricity sales amount prediction technique |
| CN109345007A (en) * | 2018-09-13 | 2019-02-15 | 中国石油大学(华东) | A kind of Favorable Reservoir development area prediction technique based on XGBoost feature selecting |
-
2019
- 2019-04-22 CN CN201910323887.0A patent/CN110033053A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105718932A (en) * | 2016-01-20 | 2016-06-29 | 中国矿业大学 | Colorful image classification method based on fruit fly optimization algorithm and smooth twinborn support vector machine and system thereof |
| CN109002918A (en) * | 2018-07-16 | 2018-12-14 | 国网浙江省电力有限公司经济技术研究院 | Based on drosophila optimization algorithm-support vector machines electricity sales amount prediction technique |
| CN109345007A (en) * | 2018-09-13 | 2019-02-15 | 中国石油大学(华东) | A kind of Favorable Reservoir development area prediction technique based on XGBoost feature selecting |
Non-Patent Citations (2)
| Title |
|---|
| 崔东文;金波;: "果蝇优化算法与支持向量机在年径流预测中的应用", 人民珠江, no. 02 * |
| 张维涛;: "果蝇优化SVM模型对有机化合物熔点的预测", 广州化工, no. 05 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109611087B (en) | Volcanic oil reservoir parameter intelligent prediction method and system | |
| US11852778B2 (en) | Static engine and neural network for a cognitive reservoir system | |
| CN107290782B (en) | Reservoir porosity, water saturation and shale content parameter Simultaneous Inversion new method | |
| CN109800863B (en) | Logging phase identification method based on fuzzy theory and neural network | |
| CN110346831B (en) | Intelligent seismic fluid identification method based on random forest algorithm | |
| CN106372748A (en) | Hard-rock tunnel boring machine boring efficiency prediction method | |
| CN106529667A (en) | Logging facies identification and analysis method based on fuzzy depth learning in big data environment | |
| CN110674841A (en) | Logging curve identification method based on clustering algorithm | |
| CN107688201A (en) | Based on RBM earthquake prestack signal clustering methods | |
| Bagheripour et al. | Fuzzy ruling between core porosity and petrophysical logs: Subtractive clustering vs. genetic algorithm–pattern search | |
| CN109345007A (en) | A kind of Favorable Reservoir development area prediction technique based on XGBoost feature selecting | |
| KR102155101B1 (en) | System and method for predicting ground layer information, and a recording medium having computer readable program for executing the method | |
| CN113902861A (en) | Three-dimensional geological modeling method based on machine learning | |
| CN107368921A (en) | Track traffic scheme comparison method based on 3DGIS+BIM technologies | |
| CN109613623A (en) | A kind of lithology prediction method based on residual error network | |
| CN115494560B (en) | Novel modeling method based on combination of CT and drilling data | |
| CN111027249B (en) | Machine learning-based inter-well connectivity evaluation method | |
| CN107316341A (en) | A kind of Multiple-Point Geostatistics facies modelization method | |
| CN103628816A (en) | Seismic attribute guidance horizontal-well-steering method and device thereof | |
| CN110988997A (en) | Hydrocarbon source rock three-dimensional space distribution quantitative prediction technology based on machine learning | |
| CN114387411A (en) | Three-dimensional geological model precision evaluation method based on variable weight theory membership function | |
| CN112861890A (en) | Reservoir evaluation model construction method and reservoir identification method | |
| CN114117898A (en) | While-drilling gamma logging forward modeling method based on machine learning algorithm | |
| CN110033053A (en) | Optimize the Favorable Reservoir prediction technique of twin support vector machines based on improvement drosophila | |
| Han et al. | Prediction of igneous lithology and lithofacies based on ensemble learning with data optimization |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190719 |