CN109918713B - Method for automatically acquiring gene knowledge of product concept design - Google Patents
Method for automatically acquiring gene knowledge of product concept design Download PDFInfo
- Publication number
- CN109918713B CN109918713B CN201910065147.1A CN201910065147A CN109918713B CN 109918713 B CN109918713 B CN 109918713B CN 201910065147 A CN201910065147 A CN 201910065147A CN 109918713 B CN109918713 B CN 109918713B
- Authority
- CN
- China
- Prior art keywords
- product
- pesticide spraying
- solid pesticide
- target product
- function
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a gene knowledge automatic acquisition method for product concept design, which comprises the following steps: analyzing function and behavior information of a target product by adopting a first-order predicate logic method, and decomposing the function information into function elements by utilizing a decomposition and reconstruction principle; secondly, finding out product examples in the same family with the target product from a product example library to form a sample set; thirdly, aggregating the samples into K classes, setting the value of K according to the type of the target product, and calculating Euclidean distances between the target product and the center points of all classes, wherein the class where the center point with the minimum distance is located is the class to which the target product belongs; extracting attribute information related to the functional elements from the sample of the class to which the target product belongs; fifthly, selecting m attributes with larger correlation with the functional elements by adopting a K-fold cross validation method; and sixthly, storing the selected m attributes and behavior information into a product gene library in a coding chain form to obtain the genetic knowledge of the target product.
Description
Technical Field
The invention belongs to the technical field of acquisition of product genetic knowledge, and particularly relates to an automatic acquisition method of gene knowledge of product concept design.
Background
The concept design is a key stage for determining the product quality and the product innovation, the stage has the least constraint on designers and is the stage which can reflect the experience, intelligence and creativity of the designers in the whole design process of the product; the stage can determine more than 80% of product cost, and the decision errors generated in the stage need to be compensated by tens of times or even hundreds of times in the later stage of design.
However, the conceptual design phase has the limitations of information, uncertainty, multidisciplinary, and a large number of design alternatives. These characteristics make the retrieval and sharing of design information difficult, the feasibility of design schemes derived from fuzzy limited information low, and the design cycle long. Therefore, there is a strong need to develop a method for determining key information in the concept design process and representing the key information in a uniform manner.
To this end, researchers have introduced the concept of "product genes," which are defined as a standardized set of information that determines key characteristics of a product, has a significant impact on the concept design process, and can be passed between parent and child products. The genetic-Based Design (GBD) method is also proved to be an effective method for developing product innovation schemes, and has the advantages of reducing the blindness of innovation Design, increasing the space of Design schemes and improving the feasibility of the schemes. Therefore, GBD is becoming more widely used in product innovation design.
The definition and acquisition of product genes is the premise and basis for the study of GBD. Currently, most of the research on the acquisition of product genes is limited to the structural aspect of the acquisition process, and an effective implementation path or method for acquiring the product genes is lacked.
Disclosure of Invention
In view of the above, the present invention provides a method for automatically acquiring genetic knowledge of product concept design, which can automatically acquire key knowledge influencing the concept design process.
The technical scheme for realizing the invention is as follows:
a gene knowledge automatic acquisition method for product concept design comprises the following steps:
analyzing function and behavior information of a Target Product (TP) by adopting a first-order predicate logic method, and then decomposing the function information into function elements by utilizing a Principle of Decomposition and Reconstruction (PDR);
step two, finding out product examples of the same family as the target product from the product example library to form a sample set;
thirdly, aggregating the samples into K classes, setting the value of K according to the type of the target product, and calculating Euclidean distances between the TP and the center points of the classes, wherein the class where the center point with the minimum distance is located is the class to which the TP belongs;
extracting attribute information related to the functional element from a sample of a class to which the TP belongs;
selecting m attributes with larger correlation with the functional elements by adopting a K-fold cross validation method, wherein m is a set value;
and step six, storing the m attributes selected in the step five and the behavior information obtained in the step one into a product gene library in the form of a coding chain, so as to obtain the genetic knowledge of the target product.
Further, in the third step, clustering is performed by using a K-means clustering algorithm.
Further, in the fourth step, the attribute information is extracted by using an Optimal First-Order-Incremental Selection (OFOIS) method based on the Max-independence and Min-Redundancy (mRMR) standard.
Has the advantages that:
(1) the method realizes the automatic acquisition of product genes, and provides a feasible and automatic method for acquiring key knowledge in the concept design process.
(2) The method makes the key knowledge in the concept design process definite, provides a unified representation method for the key knowledge, provides a basis for GBD, solves the problems of blind innovation design and low feasibility of design schemes caused by the limitation, ambiguity, uncertainty and the like of the information in the concept design process, and has important significance for improving the product design efficiency.
(3) The method realizes the acquisition of key design knowledge, and has important significance for saving the search time of the design knowledge and improving the design efficiency. In addition, the feasibility of the design scheme obtained based on the clarified key knowledge is effectively improved, and the probability that the design knowledge is referenced and reused in the subsequent design is increased, which has important significance for improving the reuse rate of the design knowledge.
(4) In the method, a K-means clustering algorithm is selected for sample clustering, a clustering result can be quickly obtained, and the algorithm is not only suitable for conventional data objects, but also suitable for big data, so that the whole algorithm has strong applicability.
(5) The OFOIS algorithm based on the mRMR has the classification effect equivalent to that of a classification method with the maximum dependency standard, the classification error rate of the algorithm is low, and the characteristic classification can be rapidly and accurately realized.
Drawings
FIG. 1 is a conceptual design process model of the present invention.
FIG. 2 is a diagram showing the gene extraction process of the product of the present invention.
FIG. 3 is a diagram showing a gene acquisition method of the product of the present invention.
FIG. 4 is a schematic structural diagram of a housing of an aircraft for spraying agricultural chemicals according to an embodiment of the present invention.
FIG. 5 is a table of some exemplary attribute parameters according to the present invention.
Fig. 6 is a two-dimensional visual clustering result graph according to an embodiment of the present invention.
FIG. 7 is a diagram of a clustering result of a portion of samples according to an embodiment of the present invention.
Fig. 8 is a graph of iteration number and cluster center in the embodiment of the present invention.
FIG. 9 is a partial index diagram corresponding to the functional unit according to the embodiment of the present invention.
Detailed Description
The invention is described in detail below by way of example with reference to the accompanying drawings.
The invention provides a gene knowledge automatic acquisition method for product concept design, which comprises the steps of modeling through a concept design process and analyzing key elements in the concept design process. Then, simulating the biological character expression process, extracting product genes from the conceptual design process model, researching the product gene coding method, and providing a product gene coding method based on functional elements.
The concept design process is a process of abstracting a target function from design requirements and expressing the target function as a structural scheme by adopting a certain technical means. Thus, the conceptual design process may be represented simply as the model shown in FIG. 1.
The function in the conceptual design process model is analogized to biological traits and the structural scheme is analogized to proteins. Since the biological trait is ultimately converted into a protein by biological gene expression, the conversion process from a functional to a structural scheme can also be considered to be achieved by the expression process of the product gene in the conceptual design. The invention expresses a conceptual design process as a function-behavior-attribute-structure scheme (FBAS) model on the basis of the traditional function-behavior-structure (FBS), wherein attributes refer to physical attributes influencing the essential characteristics of a product and geometric attributes playing a limiting role in the overall layout of the product. Therefore, the product gene extracted from FBAS can be defined as a double-stranded structure with behavior and key attributes as bases and the relationship between the two as hydrogen bonds, as shown in fig. 2. Among them, hydrogen bonds only play a role of linkage.
In order to find a unified expression method of product genes, the invention proposes a product gene coding method based on functional elements, similar to the biological gene coding method, as shown in table 1.
TABLE 1 functional element-based product Gene coding
Wherein, FiIs the ith functional element, Start is the promoter, End is the terminator, N is the gene address, ByIs a reaction of with FiCorresponding behavior, AmIs referred to as influencing FiThe expressed key attribute information mainly comprises physical attributes and geometric attributes.
The method for obtaining the product gene provided by the invention is described below by combining an embodiment of solid pesticide spraying Aircraft Shell (AS) design.
The AS is a device for forming the shape of the aircraft, containing the solid pesticide and connecting other parts of the aircraft. The device is capable of withstanding loads from both inside and outside during flight and its shape and surface quality have a significant impact on the amount of drag and friction encountered during flight. In addition, the shape of the AS is determinative of the interior space of the aircraft and the amount of solid pesticide contained therein. The structure scheme of the AS is shown in figure 4.
The main parameters of the AS are shown in table 2.
TABLE 2 Shell Attribute information Table
As shown in FIG. 3, the process of obtaining the product gene of the embodiment of the present invention is as follows:
according to the action principle of the AS, a first-order predicate logic SV (P) O method is adopted to extract functions and action verbs of the AS, S represents a subject, V represents the action verbs, O represents an action object, and P represents V or a complement of O, and the functions are decomposed into functional elements by adopting a decomposition and reconstruction Principle (PDR) technology.
TABLE 3 Main function and behavioral verbs of AS
As can be seen from table 3, the main function of the aircraft is "forming the outer shape of the aircraft", and the main behavioral verbs include "bearing", "influencing", "determining", and "connecting".
The PDR technology is adopted to carry out function decomposition on the main functions of the AS, and the obtained function elements of the AS are shown in the table 4.
TABLE 4 function element table of AS
And step two, selecting 99 instances which are in the same family with the AS from the product instance library to form a sample set. These samples have the same attribute type AS the AS, but the corresponding parameter values are not exactly the same AS the AS. Let sample set be D ═ x0,x1,…,x98Each sample can be represented as xi＝(xi0,xi1,…,xij…,xim) Wherein x isijM is 15 for the attribute information contained in the sample.
The partial parameter values of the sample set are shown in fig. 5.
And step three, clustering the sample set in the step (2) by adopting a K-means clustering algorithm, and judging the class of the target product TP.
In clustering, the number of clusters, i.e. the value of k in the algorithm, needs to be determined first. In general, the shape of an aircraft shell can be divided into three types: 1) cone-cylinder-skirt (cone angle 1 ≠ 0, cone angle 2 ≠ 0, cone angle 3 ≠ 0); 2) double cylinder (cone angle 1 ≠ 0, cone angle 2 ≠ cone angle 3 ≠ 0); 3) three cones (cone angle 1 ≠ 0, cone angle 2 ≠ 0, cone angle 3 ≠ 0). Therefore, in the present embodiment, the value of the variable k is set to 3. The clustering algorithm was implemented using python3.6, and the results after clustering are shown in fig. 6, 7 and 8. As can be seen from the figure, the algorithm reaches the end condition when the iteration reaches the 16 th time, stops the iteration, and outputs the samples contained in each class and the center point of the class. The number of samples contained in classes 1-3 after the iteration was 27,48 and 24, respectively.
After the cluster sample and the cluster center result are obtained, the class of the target product AS needs to be judged by using the Euclidean distance formula.
Where ρ (X, U) represents the Euclidean distance between X and U, XiThe ith attribute column representing the target product TP,
representing the sum x in each class centeriThe corresponding column vector.
Through calculation, the AS is known to belong to the second category: double column shape.
And step four, taking 48 samples in TP and class 2 as research objects, and sequentially selecting the attributes related to each functional element from the attribute parameters of the sample set by adopting optimal first-order incremental selection (OFOIS) based on mRMR.
The basic idea of the method is to select the features with larger class correlation according to the maximum correlation standard (formula 2), and then calculate the features selected in the maximum correlation standard (formula 3) by using the minimum redundancy standard (formula 3) to remove redundant features.
Wherein D (S, f)λ) Representing a set of attributes S and a function element fλCorrelation between, xiIs an element in the attribute set, S is the number of elements in the set S, I (x)i；fλ) Denotes xiAnd fλThe mutual information between them.
Where R (S) represents the redundancy between elements in the attribute set S, xi,xjIs the element in the attribute set S, | S | is the number of elements in the set S, I (x)i；xj) Denotes xiAnd xjThe mutual information between them.
maxφ(D-R)＝D-R (4)
Wherein phi (D-R) represents the difference between the maximum correlation D and the minimum redundancy R, and is a functional expression of the mRMR standard.
OFOIS refers to the selection of only one feature at a time from all features to add to a feature set based on the mRMR criteria. OFOIS is an algorithm equivalent to maximum Dependency (Max-Dependency) that can advance features quickly and accuratelyThe lines are classified, and the error rate of classification is very small. Suppose the total feature set is X, and we have m-1 and some function element fλRelevant features from { X-Sm-1Selecting the mth feature associated with the function element in the feature set, if it is satisfied
Wherein the selected set of m-1 features is Sm-1，xiIs a set Sm-1A certain element of (1), xjIs a set { X-Sm-1I (x; y) represents mutual information of x and y.
In this example, the key attribute associated with each function element needs to be selected, and the index value corresponding to the function element is shown in fig. 9. In the present invention, features are classified into two categories according to the relationship between the features and the functional elements: associated with a functional element and not associated with a functional element. Therefore, the OFOIS can be used for feature selection. The results obtained when implementing the algorithm in python3.6 programming language are shown in table 5, where the attributes are listed only with the corresponding subscripts.
TABLE 5 OFOIS Algorithm results Table
Fifthly, selecting attributes with large correlation with the functional elements from the attribute sequence obtained in the step (4) by adopting an eight-fold cross verification method;
as shown in table 5, an order of magnitude of correlation with each functional element is obtained using the OFOIS algorithm, but for each functional element, all features are listed with no goal of selecting the feature set with the greatest correlation. For this purpose, this embodiment further selects the attribute ranking obtained in table 5 by using 8-fold cross-validation method, so as to obtain some key attributes related to the functional element, as shown in table 6.
TABLE 6 Key Attribute information Table
And step six, storing the key attribute information obtained in the step (5) and the verb information obtained in the step (1) into a product gene library in the form of a product gene coding chain.
These attributes together with the behavioral verbs derived in step (1) constitute product genes and are stored in the product gene library in the form of product gene coding chains, as shown in table 7.
TABLE 7 product Gene Bank Table of AS
In summary, the above description is only a preferred embodiment of the present invention, and is not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (2)
1. An automatic gene knowledge acquisition method for product concept design is used for designing a shell of a solid pesticide spraying aircraft, and is characterized by comprising the following steps:
analyzing function and behavior information of the solid pesticide spraying aircraft shell by adopting a first-order predicate logic method according to an action principle of the solid pesticide spraying aircraft shell, and then decomposing the function information into function elements by utilizing a decomposition and reconstruction principle, wherein the function elements comprise various loads, influence resistance, internal space and drug amount determination and other parts connected with the aircraft;
step two, finding out a product example which is in the same family as the solid pesticide spraying aircraft shell from a product example library to form a sample set; the sample set is D ═ x0,x1,…,x98Each sample is denoted xi＝(xi1,…,xij…,xim)，Wherein x isijThe attribute information contained in the sample, m is 15, and the attribute information is shown in the following table;
thirdly, collecting the samples into K classes, setting the value of K according to the type of the solid pesticide spraying aircraft shell, and calculating Euclidean distances between the solid pesticide spraying aircraft shell and central points of the classes, wherein the class where the central point with the minimum distance is located is the class to which the solid pesticide spraying aircraft shell belongs;
extracting attribute information related to the functional elements from a sample of the type of the solid pesticide spraying aircraft shell by using an optimal first-order increment selection method based on maximum correlation and minimum redundancy standard;
selecting m attributes with larger correlation with the functional elements by adopting a K-fold cross validation method, wherein m is a set value;
and step six, storing the m attributes selected in the step five and the behavior information obtained in the step one into a product gene library of the solid pesticide spraying aircraft shell in a coding chain mode, and obtaining the genetic knowledge of the solid pesticide spraying aircraft shell.
2. The method for automatically acquiring genetic knowledge for conceptual design of products according to claim 1, wherein in step three, clustering is performed using a K-means clustering algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910065147.1A CN109918713B (en) | 2019-01-23 | 2019-01-23 | Method for automatically acquiring gene knowledge of product concept design |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910065147.1A CN109918713B (en) | 2019-01-23 | 2019-01-23 | Method for automatically acquiring gene knowledge of product concept design |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109918713A CN109918713A (en) | 2019-06-21 |
| CN109918713B true CN109918713B (en) | 2021-06-01 |
Family
ID=66960689
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910065147.1A Active CN109918713B (en) | 2019-01-23 | 2019-01-23 | Method for automatically acquiring gene knowledge of product concept design |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109918713B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110534155A (en) * | 2019-08-29 | 2019-12-03 | 合肥工业大学 | Fast packing formula Gene Selection Method based on maximal correlation minimal redundancy |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09160987A (en) * | 1995-12-13 | 1997-06-20 | Hitachi Tohoku Software Kk | Method for searching for optimum solution |
| CN101364244A (en) * | 2008-08-07 | 2009-02-11 | 同济大学 | Product gene modelling approach based on behavior semantic network identity model |
| CN101777087A (en) * | 2010-01-26 | 2010-07-14 | 同济大学 | Product function gene modeling system and method |
| CN104572590A (en) * | 2015-01-27 | 2015-04-29 | 大连大学 | Product structure design structure matrix partitioning method |
| CN105956301A (en) * | 2016-05-10 | 2016-09-21 | 北京理工大学 | Function-concept-decision model-based reconfigurable machine tool configuration design method |
| CN107944184A (en) * | 2017-09-04 | 2018-04-20 | 北京理工大学 | With reference to life system theory and the Method of Conceptual Design of FBS |
| CN108846259A (en) * | 2018-04-26 | 2018-11-20 | 河南师范大学 | A kind of gene sorting method and system based on cluster and random forests algorithm |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070181743A1 (en) * | 2006-02-08 | 2007-08-09 | Lockheed Martin Corporation | Method for streamline traced external compression inlet |
-
2019
- 2019-01-23 CN CN201910065147.1A patent/CN109918713B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09160987A (en) * | 1995-12-13 | 1997-06-20 | Hitachi Tohoku Software Kk | Method for searching for optimum solution |
| CN101364244A (en) * | 2008-08-07 | 2009-02-11 | 同济大学 | Product gene modelling approach based on behavior semantic network identity model |
| CN101777087A (en) * | 2010-01-26 | 2010-07-14 | 同济大学 | Product function gene modeling system and method |
| CN104572590A (en) * | 2015-01-27 | 2015-04-29 | 大连大学 | Product structure design structure matrix partitioning method |
| CN105956301A (en) * | 2016-05-10 | 2016-09-21 | 北京理工大学 | Function-concept-decision model-based reconfigurable machine tool configuration design method |
| CN107944184A (en) * | 2017-09-04 | 2018-04-20 | 北京理工大学 | With reference to life system theory and the Method of Conceptual Design of FBS |
| CN108846259A (en) * | 2018-04-26 | 2018-11-20 | 河南师范大学 | A kind of gene sorting method and system based on cluster and random forests algorithm |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109918713A (en) | 2019-06-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Amer et al. | A review of modularization techniques in artificial neural networks | |
| Abbasi-ghalehtaki et al. | Fuzzy evolutionary cellular learning automata model for text summarization | |
| Bi et al. | Improving image-based plant disease classification with generative adversarial network under limited training set | |
| WO2015161338A1 (en) | Ontology aligner method, semantic matching method and apparatus | |
| US20100005051A1 (en) | System and method for inferring a network of associations | |
| JP2014059754A (en) | Information processing system, and information processing method | |
| Beaufils et al. | Stochastic approach to worldwide language classification: the signals and the noise towards long-range exploration | |
| CN109918713B (en) | Method for automatically acquiring gene knowledge of product concept design | |
| Li et al. | pyDarwin: a machine learning enhanced automated nonlinear mixed‐effect model selection toolbox | |
| Pauletic et al. | An overview of clustering models with an application to document clustering | |
| Evans | Population-based ensemble learning with tree structures for classification | |
| Lamba et al. | Ranking of classification algorithm in breast Cancer based on estrogen receptor using MCDM technique | |
| CN116543832A (en) | disease-miRNA relationship prediction method, model and application based on multi-scale hypergraph convolution | |
| CN115661546A (en) | Multi-objective optimization classification method based on feature selection and classifier joint design | |
| WO2020167316A1 (en) | Machine learning-based generative design for process planning | |
| Babaiha et al. | Rationalism in the face of GPT hypes: Benchmarking the output of large language models against human expert-curated biomedical knowledge graphs | |
| KR102187586B1 (en) | Data processing apparatus and method for discovering new drug candidates | |
| Rifaioglu et al. | Multi-task deep neural networks in automated protein function prediction | |
| Rodrêguez-Fdez et al. | A genetic fuzzy system for large-scale regression | |
| CN117951457B (en) | Carbon emission reduction accounting method and system for carbon negative product based on biochar | |
| Liu | Extracting Rules from Trained Machine Learning Models with Applications in Bioinformatics | |
| CN109739977B (en) | Book concept front-rear order relation extraction method based on unsupervised learning | |
| Wan | Novel hierarchical feature selection methods for classification and their application to datasets of ageing-related genes | |
| de Campos | Optimization strategy of neural networks based on rational agents | |
| Khalilipour et al. | Intelligent Model Management based on Textual and Structural Extraction-An Exploratory Study |
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 |