- Open Access
Jacquard pattern optimizing in weft knitted fabrics via interactive genetic algorithm
© Hadjianfar et al.; licensee Springer 2014
- Received: 6 May 2014
- Accepted: 9 September 2014
- Published: 15 October 2014
A genetic algorithm is a method to respond to troubles that are indissoluble by common methods and must be utilized to try and fault method. It is difficult to appraise all of responses if there are many answers. Algorithm genetic can contain a large vast of responses and find the best of them by receiving feedbacks from problems. Several designs with different colors can be done in weft knitted Jacquard designing system. However, many patterns might not have enough attractiveness and beauty. The choice of interesting and stylish patterns of the huge set of designs according to customer judgment is difficult. An interactive genetic algorithm that received necessary feedbacks from the user, can be used in design optimization and choosing ideal patterns. In this paper a software has been constructed to optimize jacquard pattern in weft knitted fabrics based on interactive genetic algorithm.
- Genetic algorithm
- Weft knitted
Design methodology and procedure are studied, dissected and implemented in different contexts and can include principles from various disciplines. Unlike scientific inquiry, to which design methodology has sometimes been compared, design is about innovation rather than an interrogation into the nature of what exists (Parsons and Campbell 2004).
Mass customization and automated custom clothing have been regarded as promising methods in the apparel industry to create well-fitting clothing for clients. Nevertheless, custom patternmaking software cannot generate custom clothing with the perfect fit and the system cannot accomplish effectively for some customers (Song and Ashdown 2012). Song and Ashdown (2012) built up a set of basic pants patterns optimized for three lower body shape groups, and tried.
Optimizing problems can be warranted by different methods which can be split into two major groups; deterministic methods and stochastic methods. Among the latter group genetic algorithm is of more importance because of its ease in use of computers (Chamber 1995).
A genetic algorithm is a direction of optimization which is animated by the environment in which supplementary natural mechanisms such as cross over; mutation and survival of the fittest are used for machine learning. This method is widely applied in the optimization and classification problems (Goldberg 1989).
Several designs with different colors can be done in weft knitted Jacquard designing system. However, many patterns might not have enough attractiveness and beauty and opting patterns among a wide domain of customers’ taste can be a large challenge. To resolve such a challenge an optimization and designing a scheme founded on artificial intelligence can be applied. In this method a designing system based on customer’s preference and market demand is devised which progressively connects with users and gather feedbacks and perform acceptable results to users.
Kim and Cho (2000) searched on the interactive genetic algorithm used in the manner in which a database that had designing elements such as cloth body, collar, sleeve and their colors was applied which were preserved in 3D fashions.
Gu et al. (2006) used interactive genetic algorithm in designing 3D cartoon characters in which parameters such as nose, ears, eyes, mouth and shape of head were taken into account.
Genetic algorithm in designing carpet plan and woven fabrics was used (Peivandi et al. 2007a). Also interactive genetic algorithm was used for optimizing the designation of woven fabrics and similar to former research databases containing preliminary supplies were used for initializing optimization process. In this method two databases were utilized; one is connected to primary patterns and another is related to colors (Peivandi et al. 2007b). In this paper a software has been constructed to optimize jacquard pattern in weft knitted fabrics based on interactive genetic algorithm.
Like chromosomes that change in nature because of genes change, in genetic algorithm these elements are also changing continuously in order to become more complete and powerful populations. Chromosomes size and number of genes are related to the type of problem; in fact genes are true determiners of required problem optimization variables. Chromosome fitness determines the efficiency and functionality of the ones for solving required problem. In such problems that are related to peoples tastes and feelings direct fitness determination is used instead of its function by the user which are called interactive genetic algorithm. To train the following generations of chromosomes populations, the crossover operator on chromosomes is used which are in different methods; single-point crossover, two points-point crossover, cut & splice, uniform crossover (UX) & half uniform crossover (HUX) and crossover for orders chromosomes (Holland 1975). Another parameter which is effective on chromosomes is a mutation which can change some genes in a low rate accidentally. Probably for this operator is normally considered very low, about 0.001 (Mitchell 1999). A mutation in genetic algorithm is used to prevent the local convergence of the algorithm. Some of the best chromosomes in each generation are transferred to next one without any changes. Using elite chromosome traits in genetic algorithm causes an improvement in its performing trend (Mitchell 1999). Evolutionary algorithms use different evolutionary simulations in three stages of gene-chromosomes, person and generation. Genetic algorithm uses evolutionary simulation in a gene and chromosome level in which the population is consisted of chromosomes which are the same size arguments. Producing new generation is generally caused by chromosome linkage and partially by mutation.
An interactive genetic algorithm is similar to genetic algorithm; however there is a slight difference that in the first one at that place is no fitness function and the amount of fitness for each chromosome is seen by the user. Interactive genetic algorithm can communicate with users and consequently be affected by user feeling and is utilized in the areas of artistry and design (Nakanishi 1996a,[b]; Ohsaki et al. 1998; Takagi 1998). Software for optimization of designs is developed by MATLAB and for this; different portions of it such as Graphical User Interface Builder (GUIED) and Genetic Algorithm Tod are used.
The devised software has six sections. Patterns and colors made by the software are demonstrated to the user in the main piece which takes user score. In other parts user can observe the existing patterns in pattern database and create, edit or delete desired pattern and impose it to the database. User can also make out the same for the colors in a component of software and finally introduce the limit of knitting machine in pattern and color. Each pattern is formed by a 3D matrix in which each component represents a loop, the number of rows represents the number of course and the number of columns shows the number of wale in the pattern. The first dimension of matrix shows the amount of red, the second one shows the same for green and the third one shows the same for blue. In training to create a new pattern user first sees the number of course and wale and color of the design, then designs his own design and add it to the database.
Developing fabrics with the desired texture by the means of knitting machines is limited due to the limitation of number of feeders and increase of colors which causes a reduce in production. Since it is probable that all the patterns in the database are not capable of being produced by jacquard knitting machine, patterns database is modified by a user based on imposing limitations on the number of machine colors and patterns with more colors than machine restriction will be eliminated.
2.1 Pattern optimization
Design chromosome gene values
2.2 Color optimization
Unlike design chromosomes with the same number of genes, the number of genes for color chromosomes can change and is determined according to color variety of selected designs. Number of genes equals the number of selected design colors. Color chromosome has an auxiliary gene for determining the number of colors or in other words the number of genes which will be eliminated after the color chromosome structure is formed. The main color chromosome genes gain the required value based on the number of existing colors in color database. For example, if there are sixteen colors in color database the first gene gets a value 1 to 16 after the auxiliary gene is eliminated. The following genes cannot take the previous genes values because the design may encounter to unsolicited changes.
Color optimization has a lower rate than the design optimization which is caused by a wide variety of colors.
The devised software has an acceptable capability for design and color optimization. The user interface is powerful software in obtaining data and tastes from the user which can adjust itself to different positions and impressions. This software can create more beautiful and attractive designs and colors in the standpoint of the user during making new generations. Fitness diagrams also indicate an improvement in color and design scores.
Color optimization has a lower rate than the design optimization. Due to a wider variety of colors, their scoring range is broader, too. Consequently, more generations are necessary in this instance. Moreover, automatic scoring of repeated designs helps the software performs faster. One of the significant aspects of this software is the existence of a variable database which can make a better optimization and increase design range size.
- Chamber L: Practical Handbook of Genetic Algorithms. 1995, CRC Press, Boca RatonView ArticleGoogle Scholar
- Goldberg DE: Genetic Algorithms in Search, Optimization and Machine Learning. 1989, Addison-Wesley Publishing Co. Inc, BostonGoogle Scholar
- Gu Z, Tang MX, Frazer JH: Capturing aesthetic intention during interactive evolution. Computer-Aided Design. 2006, 38: 224-237. 10.1016/j.cad.2005.10.008.View ArticleGoogle Scholar
- Holland J: Adaptation in Natural and Artificial Systems. 1975, University of Michigan Press, MichiganGoogle Scholar
- Kim HS, Cho SB: Application of interactive genetic algorithm to fashion design. Engineering of Artificial Intelligence. 2000, 13 (6): 635-644. 10.1016/S0952-1976(00)00045-2.View ArticleGoogle Scholar
- Mitchell M: An Introduction to Genetic Algorithms. 1999, MIT Press, LondonGoogle Scholar
- Nakanishi Y: “Applying evolutionary systems to design aid system”, Proc. Of Artificial Life V (Poster Presentation). 1996Google Scholar
- Nakanishi Y: “Capturing preference into a function using interactions with a manual evolutionary design aid system”, Genetic Programming 1996 Late-Breaking Papers. 1996Google Scholar
- Ohsaki M, Takagi H, Ingu T: “Methods to reduce the human burden of interactive evolutionary computation”, Proc. Of Asia Fuzzy Systems Symposium. 1998Google Scholar
- Parsons JL, Campbell JR: Digital apparel design process: Placing a new technology into a framework for the creative design process. Clothing and Textiles Research Journal. 2004, 22 (1/2): 52-88.Google Scholar
- Peivandi P, Latifi M, Amani-Tehran M: “Genetic algorithm aided loading layout of machine-made carpet”, 6th International Conference of Textile. 2007, Isfahan University of Technology, IranGoogle Scholar
- Peivandi P, Latifi M, Amani-Tehran M, Akbari-Moayed H: “Woven pattern design using interactive algorithm genetic”, 6th International Conference of Textile. 2007, Isfahan University of Technology, IranGoogle Scholar
- Song HK, Ashdown SP: Development of Automated Custom-Made Pants Driven by Body Shape. Clothing and Textiles Research Journal. 2012, 30 (4): 315-329. 10.1177/0887302X12462058.View ArticleGoogle Scholar
- Takagi H: “Interactive evolutionary computation-cooperation of computational intelligence and human KANSEI”, Proc. Of Int. Conference on Soft Computing. 1998Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.