Papers on Interactive Genetic Algorithm to Fashion Design

1 Introduction

A designer should always keep the end user in mind when designing products. After all, the cease user will generally be in contact with the product. Affordance based design (ABD), introduced by Maier & Fadel (Reference Maier and Fadel2001), describes the interactions betwixt the end user and the antiquity (artifact–user affordances, AUAs) using the concept of affordance. Although there are many product evolution methods, some methods are ameliorate platforms for user involvement than others. Affordance based design has the advantage of using the concept of affordance, which is used to describe the possible means to interact with or use the product. This suggests that users can evaluate the affordances of a production by observing it. This newspaper explores how end-user feedback can be captured through the evaluation of perceived affordance quality and, in plow, be used to evolve the class of products toward a better satisfaction of these affordances.

1.ane Affordance based blueprint; the affordance concept

The term affordance emerged from the field of perceptual psychology (Gibson Reference Gibson1979). Gibson defined it by 'The affordances of the environment are what information technology offers the animal, what information technology provides or furnishes, either for good or ill.' Information technology was created to describe what a system (e.chiliad., an antiquity) provides to another organisation (eastward.yard., a user). Norman (Reference Norman1988) then extended the term to aid in the design of everyday things, but he stopped short of incorporating the concept of affordance as fundamental to the blueprint of any antiquity. Maier & Fadel (Reference Maier and Fadel2001) introduced the concept of affordance as being fundamental to engineering design and defined information technology as a human relationship betwixt two subsystems in which potential behaviors tin can occur that would not be possible with either subsystem in isolation (Maier & Fadel Reference Maier and Fadel2009a ). Furthermore, these authors described the 5 central properties of affordances (Maier & Fadel Reference Maier and Fadel2009b ): complementarity, polarity, multiplicity, quality and form dependence.

ane.2 Product evolution by improving affordances

As described earlier, one of the properties of affordances is that of quality. Quality tells us how well a organization affords a specific use or action, in this example, based on the perception of the user. This definition of quality has been embraced by other authors (Cormier, Olewnik & Lewis Reference Cormier, Olewnik and Lewis2014; Ciavola, Wu & Gershenson Reference Ciavola, Wu and Gershenson2015; Ben Hamida et al. Reference Ben Hamida, Jankovic, Curatella, Baltay, Callot, Huet, Bocquet and Sasse2016). Products often go through many iterations (e.g., vacuum cleaners, cars, ability tools), and it can be assumed that their quality increases in every iteration. Gaffney, Maier & Fadel (Reference Gaffney, Maier and Fadel2007) plant that the evolution of these products tin be explained with an improvement in the quality of their affordances.

The quality of the affordances tin be assessed either past the designer of the artifact or past unlike types of users (e.one thousand., manufacturing, maintenance, end users, etc.) (Maier & Fadel Reference Maier and Fadel2009b ). Utilize of the input from end users would help the designer to know what users perceive to be a high-quality production. Information technology should be noted that the affordance quality assessment past users is naturally subjective. A person could concur that the sitting-ability of a chair is better than that provided past a briefcase, only could all the same use the sitting-power of the briefcase, not minding the inferior quality information technology offers. On the other hand, there could be some other person who would not tolerate the sitting-ability offered by the briefcase and would rate the affordance quality of the chair significantly college. This means that there is no right way to appraise an affordance quality every bit information technology reflects the perceptions of users.

The challenge and so is to utilize that feedback from users directly in the pattern process. It should exist noted that affordance perception is not a variable in this framework. If a user is being asked about the quality of a particular affordance, it can be assumed that the user perceives that affordance. This does non mean that users cannot perceive whatever other affordances besides the ones being suggested to them. With an appropriate interface, users could suggest that an affordance exist added to the product.

Ciavola et al. (Reference Ciavola, Wu and Gershenson2015) establish a way to integrate the concept of affordances with genetic algorithms (GAs) and make some concrete characteristics of objects change according to the evaluation of the quality of their affordances. Nonetheless, their inquiry did not make use of existent cease users; instead, they implemented a neural network that was trained by i user and mimicked input from that user.

The do good of GAs is that they allow real user assessments to exist processed directly, which means that the optimization and feedback processing is washed in one footstep. As user assessments can change over time, neural networks would need to be retrained to consider the dynamic nature of people's assessments.

Figure ane. Affordance based pattern/genetic algorithm integration.

This inquiry builds on Nguyen et al.'southward research. The ABD and GA integration (Figure 1) requires a platform where users tin be reached. In the figure, 'x' represents a blueprint solution; 'f' represents the evaluating function with an affordance quality output. To this stop, a web awarding has been created that enables users' cess of the quality of the affordances of a product, and their inclusion in the design process to evolve the product toward a more 'optimal solution'. Due to the fact that users collaborate with an optimizer (the GA), the GA becomes an interactive genetic algorithm (IGA).

The user assessment of affordance qualities is closely related to obtaining a utility office that represents the preferences from users. The utilise of preference models has been previously used in optimization (Orsborn & Cagan Reference Orsborn and Cagan2009; Reid, Frischknecht & Papalambros Reference Reid, Frischknecht and Papalambros2012), where a office that represents user preferences is adamant and so is used as an optimization objective. Preference models are non needed when an IGA is implemented; apply of an IGA is basically an alternative (more examples are provided later on) to capturing user preferences (Ren & Papalambros Reference Ren and Papalambros2011). The way in which user preference is unremarkably captured is by having users select a preferred solution out of many options (Orsborn, Cagan & Boatwright Reference Orsborn, Cagan and Boatwright2009; Reid, Gonzalez & Papalambros Reference Reid, Gonzalez and Papalambros2010; Kelly et al. Reference Kelly, Maheut, Petiot and Papalambros2011; Ren & Papalambros Reference Ren and Papalambros2011). In this paper, user preference is captured in the form of affordance quality perceptions, where the GA is in charge of evolving solutions based on the input from multiple users. The advantage of using the affordance/GA integration lies in its ability to solve multicriterion scenarios, a feature that is needed when multiple affordances can exist associated with a single product.

1.three Affordances versus requirements

Affordances should not be confused with requirements. Requirements are the objectives that the intended solution is expected to satisfy besides every bit the backdrop it must take (Pahl et al. Reference Pahl, Beitz, Feldhusen and Grote2006). At that place are many requirement classifications; the simplest classification has two types of requirements: demands and wishes (Pahl et al. Reference Pahl, Beitz, Feldhusen and Grote2006). Demands (in other words, constraints), which can have a fixed target value or a range of target values, must be fulfilled for the blueprint to be acceptable. Wishes practice not have to be fulfilled, but when evaluating design variants, those with nigh wishes fulfilled, and that better see the demands, are preferred (Pahl et al. Reference Pahl, Beitz, Feldhusen and Grote2006). Maier & Fadel (Reference Maier and Fadel2009a ) suggested that some requirements should exist interpreted in terms of affordances, to afterward find solutions that offering those affordances, which in plow fulfill these specific requirements. Affordances are not requirements in the same mode that functions (transformations of inputs to outputs) are non requirements. Even though functions can hands be linked to requirements (run across Figure 6.6 in Pahl et al. (Reference Pahl, Beitz, Feldhusen and Grote2006)), functions are considered to be the middle pace between the problem definition (through abstraction of the tasks that the production has to fulfill) and the solution (structuring of the concrete product through working principles). The same logic can be practical to ABD. Affordances tin be a manner to interpret certain requirements (abstracting requirements in terms of interactions between design entities), for which solutions accept to be institute in order to fulfill those requirements. An important difference from function is that affordances are perceived and therefore are related to the physical characteristics of the artifact. Information technology should be noted as well that different users may identify a multiplicity of affordances once they use the antiquity. These are, nigh of the fourth dimension, non requirements, although they may become and so if the designer believes that there is an added value in highlighting that interaction.

The advantage of using affordances is that they can exist hands assessed by users (due to the quality holding) to determine, for example, optimal product forms. Of course, non all affordances tin be assessed by users. For example, users would non be able to assess the affordances related to the rack and pinion system of a steering wheel to turn the car, simply because they cannot 'perceive' this system when using a steering wheel. Nevertheless, users would be able to assess the turn-ability of a steering wheel, which encompasses other affordances. For example, users might be able to perceive a departure between an assisted rack and pinion system (power steering) and a non-assisted system. Even though they do not directly assess the artifact–antiquity affordances (AAAs), the effects of these systems can be assessed through the plow-ability affordance (an AUA). The benefit of using the ABD framework in this scenario is the power to synthesize the results of users' evaluations of the product through the evaluation of AUAs.

1.iv Interactive genetic algorithms

Interactive genetic algorithms are GAs where the evaluating function is substituted by human users that collaborate with the GA through an interface.

An IGA platform could optimize both qualitative and quantitative criteria at the same fourth dimension (Brintrup et al. Reference Brintrup, Ramsden, Takagi and Tiwari2008). This paper discusses a platform that focuses on qualitative criteria, the affordance qualities that users perceive when examining a prototype. Furthermore, through this do, by studying the human relationship between concrete characteristics and affordance quality evolution, design noesis conquering that can be used in the apotheosis design phase is extracted.

1 of the challenges of using IGAs is the large number of evaluations that need to be fabricated by users (Hsu & Chen Reference Hsu and Chen1999). As Banerjee, Quiroz & Louis (Reference Banerjee, Quiroz and Louis2008) pointed out, it is beneficial to accept multiple users make these evaluations since they diversify the results. Researchers take worked on solutions to this problem (Hsu & Chen Reference Hsu and Chen1999), only these solutions involve approximating users' input for the concepts they practise not assess. The affordance based interactive genetic algorithm (ABIGA) application solves this problem past allowing parallel evaluations with a single CPU running the GA, specifically AMGA2 (Tiwari, Fadel & Deb Reference Tiwari, Fadel and Deb2011), a GA that works with small-size populations ( ${\sim}10$ , every bit opposed to ${\sim}100$ in normal GAs).

Banerjee et al. (Reference Banerjee, Quiroz and Louis2008) implemented IGAs to evolve floorplans and widget layouts. They accomplished satisfactory results within 15–20 generations/iterations. Banerjee et al. also establish that designs evolved by a collaborative peer group were consistently rated higher on the 'originality' scale when compared with designs evolved by a single designer. This is implemented in the ABIGA framework by assuasive multiple people to rate unlike concepts. The previously mentioned authors used a collaborative framework where individual IGA sessions could connect to each other to send and retrieve concepts. This means that such setups use multiple IGAs, each user sharing one 'best' solution per generation. The users can select solutions from other users and incorporate them in their ain population.

Brintrup et al. (Reference Brintrup, Takagi, Tiwari and Ramsden2006) and Brintrup, Ramsden & Tiwari (Reference Brintrup, Ramsden and Tiwari2007) used IGAs to optimize manufactory layout designs. At that place were ii types of IGAs: a sequential and a multi-objective IGA. The sequential IGA could be set to optimize using quantitative and qualitative objectives sequentially. The multi-objective IGA considered both qualitative and quantitative objectives simultaneously. In either instance, the user was responsible for the qualitative assessments of all individuals of each generation. Their results showed faster convergence when optimizing quantitative and qualitative objectives at the aforementioned fourth dimension.

Brintrup et al. (Reference Brintrup, Ramsden, Takagi and Tiwari2008) conducted a similar experiment where they used an IGA to have users assess two subjective pattern parameters (comfort and liking of a chair). These can be translated to affordances of a chair. In Brintrup et al.'south paper, each user was responsible for evaluating all of the concepts in each experiment (all generations).

One of the limitations in previous research that implements IGAs to evolve designs is the ease of user access. In all of the research previously addressed, the computer application most probable needs to be downloaded past each of the users. This as well limits the number of users that can be reached. To solve this issue, the ABIGA application was designed to be web-based. Anyone with an Internet connection and a web browser tin employ ABIGA. This application can be accessed using desktops, laptops, tablets and smartphones.

1.5 Description of the spider web application

The ABIGA is a platform that evolves product variants using the input from end users. The spider web application allows designers to set up design problems that can and so be made available to users. The blueprint bug require the specification of the design parameters of the artifact, the minimum and maximum values that these parameters can prefer, a virtual representation of the artifact and the list of affordances that the users will evaluate. Once a design experiment has been initialized, users can access the application through a web browser, select the experiment and evaluate the affordances of the antiquity in question.

A GA, AMGA2 (Tiwari et al. Reference Tiwari, Fadel and Deb2011), is used to improve and evolve the pattern solutions available to users. The GA considers the affordances equally the objectives of the problem and the design parameters as its design variables. A design concept, or solution, is defined by its set of design parameter values. These parameters are used to depict each concept for users to see and evaluate its affordances.

The GA is therefore solving multi-objective optimization problems. The optimization problem is defined as follows:

$$\brainstorm{eqnarray}\displaystyle & & \displaystyle \text{maximize}\{f_{1}(\mathbf{ten}),f_{2}(\mathbf{x}),\ldots ,f_{k}(\mathbf{10})\}\nonumber\\ \displaystyle & & \displaystyle \text{subject to }\mathbf{x}\in S,\nonumber\finish{eqnarray}$$

where the $f_{i}(\mathbf{ten})$ represent the affordances of the artifact evaluated past users. The variable vectors $\mathbf{x}$ belong to the non-empty feasible region $S\subset R^{n}$ and represent the design variable values that define a design solution.

Figure 2. The ABIGA operation.

When the experiment is started, an initial population is created past the IGA. The ABIGA creates the solutions from the IGA population, which multiple users can interact with and evaluate (see Figure 2). Once a population is evaluated in one generation, the IGA creates a new population for the side by side generation. Improved pattern concepts are created in every generation. The stopping criterion is currently gear up by specifying the number of office evaluations. The number of IGA generations can be determined because the number of solutions in each generation is known beforehand. The stopping criterion could also be ready every bit an fault between the population fettle average and the upper bound of that score.

1.half-dozen The GA: AMGA2

Any multi-objective GA can be used as the optimizer in ABIGA. The archive based micro genetic algorithm 2 (AMGA2) is a multi-objective evolutionary algorithm (MOEA) that borrows concepts from multiple MOEAs (Tiwari et al. Reference Tiwari, Fadel, Koch and Deb2008, Reference Tiwari, Fadel and Deb2011) and was readily available to the authors. This algorithm works with a small population size and keeps an external archive of skillful solutions constitute.

The small population size (number of solutions in each GA generation) of this GA allows for a pocket-size number of users (5–x users) to evaluate and execute the optimization. The size of the population can be as low every bit two times the number of objectives (Tiwari et al. Reference Tiwari, Fadel and Deb2011). The number of users can exist large, and, as the population of a generation is evaluated, a new ane is generated and presented to the users. The small size of the population does mean that not many evaluations are needed to complete a GA iteration. The algorithm can have a reduced number of solutions in the population due to the use of an archive. An annal is created at the beginning of the run; once all solutions of the archive are evaluated, the GA starts to iterate with a small population size and eventually replaces better solutions in the archive.

The AMGA2 randomly regenerates the initial population of solutions using Latin hypercube sampling (Loh Reference Loh1996) along with unbiased Knuth shuffling. A slightly modified version of differential evolution (DE) (Kukkonen & Lampinen Reference Kukkonen and Lampinen2005) is used as the crossover operator, which allows existent variables to exist used. In other words, DE allows the design variables to be encoded with real continuous values. The probability of mutation in AMGA2 is dynamic – it does non demand to be tuned for specific problems; information technology is based on the rank of the parent solutions which can change throughout the optimization run.

1.7 Development of the spider web app

The ABIGA is written using Eclipse Luna (2014), an integrated development environment (IDE). Many programming languages are involved in the evolution of ABIGA. Java (2015) is the main linguistic communication that controls nigh of the information. JavaServer pages are used to show data to the users, and utilize HTML to attain that function. JavaScript and JQuery are used to make the web pages dynamic.

The data fed to and generated by the application is saved in a database (coded in MySQL). The Google App Engine (2015) platform is used to deploy the application. All of the services used in the awarding are provided by Google (database, datastore, Java runtime environment).

1.7.1 The model-view-controller architecture

The model-view-controller (MVC) architecture was followed to construction the code of the spider web application. This structure helps to separate the code that creates and handles the information from the code that presents the data (Hall, Dark-brown & Chaikin Reference Hall, Brown and Chaikin2008), taking advantage of the strengths of different technologies. For case, JavaServer pages are skilful at presentation, even though they allow Java code to be used; their strength is in presenting data. The forcefulness of servlets is the processing of data, even though they let information technology to be presented as well. Having the code structured with the MVC architecture allows for easy code modification and improvement of the application.

Figure 3. The ABIGA user interface.

i.7.two User interface and spider web pages

The user interface shows the pattern concept to the user as well as listing suggested affordances of the concept. Sliders with a value range from $-$ three to $+$ three (seven-point Likert scale) are given for each affordance so that the user can assess the quality of those affordances. It should be noted that in the current implementation, only positive affordances are being proposed to the users and the quality rating that the users volition provide is explained in the instructions, equally shown in Figure 3.

Other scales could have been selected. The easiest calibration for users to employ would be ane with only three values, e.grand., bad, neutral and good. However, this would non exist sufficient to drive the GA, and then a larger scale was warranted. Although a scale with a large range would be meliorate for the GA, it would arrive difficult for users to sympathize and differentiate quality levels. For example, in a $-$ 10 to $+$ ten scale, users would accept bug in understanding the differences betwixt a $+$ iv and a $+$ 6. The $-$ 3 to $+$ 3 calibration was enough to drive the GA while at the same fourth dimension making it easier to assess quality.

The interface of the awarding is based on visuals only. Some affordances would require the user to affect the antiquity to effectively appraise its quality. But affordances that tin can be assessed through a visual interface can be implemented in the electric current build of the application. There are technologies that could allow users to feel the solutions through haptic feedback controllers, merely this limits the number of users that tin exist reached. For this reason, this research focuses on visual perceptions.

Figure 3 shows a screenshot of the user interface. Instructions are provided to the user on how to evaluate the affordances of the product and what their meaning related to quality is. The center of the page is where users spend nearly of their time. The left department shows a second drawing of the product, which changes co-ordinate to the solution extracted from the database. The product is placed in an environment that users are likely to see when they use it. This is to give the user some idea of the relative size of the production and its components with respect to other objects they might exist familiar with. The correct section has a list of the affordances of the product. Each affordance has a slider where users specify the quality of that affordance for the concept shown on the left. There are two buttons below the affordance list on the page. The green button sends the results for the electric current solution to the database and loads a new solution for users to evaluate. The red push sends the solution evaluations to the database and logs the user out of the experiment, which means that they practise not get whatsoever more solutions to evaluate. The bottom of the page (not shown in the effigy) contains the descriptions of all of the affordances of the product. Users are instructed to read these before they begin their evaluations and can always refer dorsum to them in instance they need to.

The evaluation page is not the simply page that users become to interact with. There are other pages that users see before they get to the evaluation page. For case, users are required to provide some information (age, sex activity, education level, name initials) in a login page, they have to select the experiment they wish to be a function of in an experiment option page and, finally, they are taken to a 'cheers' folio when finished.

Users are not expected to exist consistent in their evaluations. User evaluation depends on the mapping between the feature parameter space (the real world, which can exist described objectively) and the human psychological space. This mapping changes constantly, so even if the same feature parameters are shown to a user at different times, different evaluations may be expected. Research has been conducted that reports that this type of input noise does not bear upon IGA (Ohsaki, Takagi & Ohya Reference Ohsaki, Takagi and Ohya1998) convergence. The reason is that the IGA is non expected to converge on a point but rather an area in the solution infinite.

1.7.three How a database is used to freeze the GA

Different many IGAs, ABIGA allows parallel user evaluation using a single GA running on the server. Some changes had to exist made to the way in which the IGA works. Genetic algorithms, for the most function, make solution evaluations in sequence. This tin become an upshot for IGAs, considering it would mean that if multiple users are used equally evaluators, they would have to evaluate one at a time, i.e., each would have to wait for other users to finish evaluating their solutions.

This claiming was overcome in ABIGA using Google's Datastore technology. The Datastore is a schemaless NoSQL scalable storage service. The Datastore allows the storage of data objects (such equally Coffee classes that hold data). When a new generation of solutions is created, the GA data are saved in the Datastore and the generation solutions are saved in a database. Because of this, the GA does not need to run while all of the solutions are existence extracted from the database to be evaluated. This also means that multiple users can evaluate solutions at the aforementioned time, significantly improving the total user evaluation time. One time all solutions for a generation are evaluated, the data are retrieved from the Datastore and used by the IGA to generate a new set of solutions.

3 Experimental results

The ABIGA stores a lot of data for each blueprint experiment. Every concept generated by the GA is stored in the database. This includes the design parameter values and the affordance evaluations of each solution. All of this information tin can exist queried from the database for analysis while the experiment is running or after it is completed.

3.1 Product evolution

To check whether the solutions in the IGA improve across the different generations, the fettle of the unabridged population can exist tracked. Since in that location are multiple objectives (number of affordances), the overall fitness of a solution tin be reduced to the sum of all of its objectives. This is done only for piece of cake visualization of the evolution of the product. This is not how the IGA operates. The IGA is a multi-objective algorithm that uses the rankings of each solution with respect to each objective to evolve solutions toward amend ranked solutions that eventually go non-dominated, and perchance Pareto. Details on how the GA works can exist found in Tiwari et al. (Reference Tiwari, Fadel, Koch and Deb2008, Reference Tiwari, Fadel and Deb2011). The graph does prove, however, the trend toward solutions that are perceived to be meliorate by the users, considering all of the affordances. Figures five and half-dozen show the average of solution fitness values beyond all generations for experiments one and 2, respectively.

Figure 5. Steering bike evolution; experiment 1.

Figure six. Steering cycle evolution; experiment 2.

The graphs bear witness 16 generations considering the showtime generation is the initial population, which also has to be evaluated by users. The maximum value possible for the overall fitness is 15, as at that place were five affordances, each of which could have a maximum value of three. The minimum value is $-$ 15 because the lowest quality value for each affordance is $-$ three.

A steady increase in the fitness of the population tin be seen after the 12th generation in experiment 1. The results of experiment ii show a similar behavior, starting from generation ten. This means that the solutions generated by the IGA were perceived to be better in quality than the initial population solutions every bit a whole. It should exist noted that this does non hateful that there were non whatsoever solutions rated equally good solutions in earlier generations. To prove how all objectives amend across generations, Figure 7 compares solutions from generation ane with the solutions in the last generation (15). Three solutions were selected for these generations: depression-, medium- and high-valued solutions.

To test whether product evolution can be attained out of risk, 2 experiments were performed using a random number generator (RNG) as the evaluator of quality affordance. The results are shown in Figures 8 and ix.

Figure 7. Affordance quality evolution; experiment 1.

Figure 8. Steering wheel development RNG; input one.

Effigy 9. Steering wheel evolution RNG; input two.

As well saving the solutions generated by the GA in the database, the best solutions are saved separately for easy access. These solutions are solutions found in the non-dominated front at the finish of the optimization run. Figures 10 and 11 show some solutions generated by the IGA on the first generation and the last generation of experiment i, respectively.

Figure 10. Subset of generation 1 solutions; experiment 1.

Figure eleven. Subset of archive solutions; experiment one.

Even when ABIGA provides a fix of optimized solutions, the designer yet needs to perform other types of evaluations that cannot be obtained through user evaluations, nearly probable because other parameters on those solutions. Instead of suggesting a specific value for each of the parameters, the designer could use value ranges obtained from the assessments of the users. To determine these ranges, the data would have to show that there are relationships betwixt the affordances and the design parameters of the product.

iii.2 Relationships between affordance quality and blueprint parameters

The being of relationships between affordances and design parameters is suggested by one of the backdrop of affordances, course dependency. This property says that the affordances in a product are dependent upon the shape or geometry or physical characteristics of objects. For case, the form of a big box does not offering hand grip-power, but if a handle is added to the sides of the box, therefore changing the geometry of the box, grip-power is now possible due to that change. Having relationships between affordance qualities (equally perceived by users) and the design parameters of the solutions would mean that the designer could select blueprint parameter values that consistently go positive ratings by users to target specific product affordances.

The affordance quality input from users is categorical; that is, the variables are detached, ranging from $-$ 3 to $+$ 3. Users were instructed that negative values meant a bad affordance quality, nil meant neutral and that positive values meant that the production had a adept affordance quality. Unlike the affordance variables, the design parameter variables are continuous. Due to this discrepancy, linear regression techniques cannot be implemented between these variables to determine whether there is any relationship between them. Instead, binary logistic regression was implemented as it is the recommended method to test relationships between categorical and continuous data.

The affordance variables were further categorized into a binary type of response. Tabular array 3 summarizes how this categorization was made. The zero-valued responses were not used in the assay.

Table 3. Binary categorization of user response

The reason for making the affordance variables binary is because designers are interested in what users consider to be 'good' solutions. If a relationship between a design variable and an affordance exists, the binary logistic regression can tell us the design parameter values that are more likely to be perceived every bit positive by the users.

The logistic regression is represented by

(1) $$\brainstorm{eqnarray}\ln \left(\frac{p}{1-p}\right)=\unicode[STIX]{x1D6FD}_{0}+\unicode[STIX]{x1D6FD}_{1}x,\cease{eqnarray}$$

where $p$ is the estimated probability that a user would positively rate an affordance given a design parameter value of $10$ . The intercept $\unicode[STIX]{x1D6FD}_{0}$ and the coefficient $\unicode[STIX]{x1D6FD}_{one}$ are given in the logistic regression results for all relationships institute. With this equation and the results of the logistic regression for two variables, the designer can determine the value of the design parameter for a desired probability of acceptance.

Table four. Binary logistic regression $p$ -value results; experiment 1

Table 5. Binary logistic regression $p$ -value results; experiment 2

The binary logistic regression assay was carried out using Minitab16 (2016), and therefore non performed past ABIGA itself. All of the solutions generated past the GA are used in the binary logistic regression tests. Tables 4 and 5 evidence the results of pattern parameters tested against all affordances for both experiments (the Hosmer–Lemeshow goodness of fit exam results are given in the Appendix). The $p$ -values of the logistic regression tests are shown in the tables for each pair of pattern parameter and affordance. The number of observations for each binary logistic test varies according to how many data points have been omitted due to not considering the zero-valued responses. The boilerplate number of observations is 180. If the $p$ -value is less than 0.05, and so there is evidence that the design parameter contributes to the prediction of the affordance quality consequence. The meaning values are highlighted in the tables.

Table half-dozen. Experiment i logistic regression coefficients $(\unicode[STIX]{x1D6FD}_{0}\mid \unicode[STIX]{x1D6FD}_{1})$

Table 7. Experiment two logistic regression coefficients $(\unicode[STIX]{x1D6FD}_{0}\mid \unicode[STIX]{x1D6FD}_{one})$

The logistic regression coefficients, $\unicode[STIX]{x1D6FD}_{0}$ and $\unicode[STIX]{x1D6FD}_{i}$ , are given in Tables 6 and seven only for the pregnant relationships. These coefficients can be plugged into Eq. (i) to mathematically describe the significant relationships betwixt affordances and blueprint parameters. An example of how this is used is provided in the post-obit section. It should be noted how the sign of $\unicode[STIX]{x1D6FD}_{1}$ defines whether the relationship is positive or negative. A positive sign indicates that the quality perception probability increases with an increase of the blueprint parameter value.

iv Give-and-take

4.1 Production evolution

Figures v and six show how the IGA is able to assemble and combine the input from users to create solutions that are perceived to be high quality by most users. Yet, this does non mean that the optimized solutions provided by the IGA can be straight transferred to product. As mentioned in a higher place, the designer, of grade, nevertheless needs to carry out other types of assay on these solutions to make sure that they run across other pattern criteria that are non represented within the affordances evaluated past the users. For example, the designer might need to perform a stress analysis on the chosen solutions to make sure that the production will not fail under usage loading. In spite of maybe obtaining design solutions that practice not meet other design criteria, the information given by the IGA is valuable as it provides design variable values that are most likely to be perceived as loftier quality by the users.

The results of the IGA are in accord with the findings of Nguyen et al. (Reference Nguyen, Guarneri, Fadel and Mata2012), which showed that the average fitness of the population across generations reached a maximum at about fifteen generations. This result is important considering their steering wheel problem was single objective. The problem solved in this paper is multi-objective. This means that an increase in the number of objectives did not increment the number of generations needed to reach loftier population fitness averages. This suggests that an increase in the number of objectives (product affordances) may not negatively touch on the number of generations that the IGA should perform to detect practiced solutions.

The results in Figures ten and 11 prove the full general trends in the blueprint of the solutions of the initial population and the best solutions saved in the archive of the IGA for experiment 1. The angle betwixt the top 2 spokes has college values in the archive than in the initial population. This makes sense as this pattern parameter allows users to come across through the steering wheel to better appreciate the gauges on the dashboard of the car. The thickness of the ring seems to be larger in the annal than in the initial population. This suggests that users are more than likely to perceive a thicker band as being of higher quality than a sparse ring, similar the starting time solution shown in Figure 10.

4.2 The bad evaluator trouble

It might seem that if a user deliberately gave solutions bad quality values, the IGA would therefore provide bad results. Due to the way in which GAs work, this is non a problem with ABIGA. To search for the best solutions, GAs combine proficient solutions (crossover functioning) to create new solutions. This ways that even if there is a user who consistently gives products bad assessments, the IGA volition about likely not be also affected past those solutions. This does non mean that all of the solutions chosen for the crossover operation are good solutions; at that place is always a pocket-sized gamble that bad solutions will be called considering the IGA will effort to explore as much of the solution infinite every bit it can. The only, and unlikely, scenario that could touch the evolution of a product in ABIGA is when the entirety of users in an experiment purposefully give bad assessments.

4.3 Pattern parameters versus affordances correlations

Equally is expected in multicriteria problems, at that place are always tradeoffs, which means that there is no perfect solution. Because of this, there is no single solution that volition be preferred by all users. Therefore, information technology does not make sense for the designer to await to utilise specific design parameter values obtained from ABIGA in their design. After all, we would not expect that a quarter degree in the angle between the top ii spokes of a steering wheel would make much of a difference to the perception of quality from users. It would make more sense if designers had ranges of values for each design parameter that they could piece of work with so that whatever values were chosen would elicit good quality perceptions from the users. Information technology turns out this is possible if the responses from the users show that there are relationships between the design parameters and the affordances of the product in question.

Equally shown earlier, using logistic regression techniques, relationships between design parameters and affordances were found. These relationships have been highlighted in Tables 4 and 5. These results suggest that designers can target specific affordances by changing the values of pattern parameters so that users perceive a loftier quality for those affordances.

Information technology should be noted that the use of a logistic regression technique to check for these relationships inherently limits how the relationships are described. The dependent variable (affordance quality) is turned into a stochastic event (skillful/bad affordance quality, as described in Table 3). This is described with a density office of cumulated probabilities ranging from naught to one. This ways that other models cannot exist used to depict these relationships, such as quadratic or cubic models. Time to come experiments could be modified, making affordance quality a continuous variable to permit for other types of models to be fitted. Still, every bit will be shown after in this section, logistic regression provides valuable information to designers.

I of the relationships found was between Turn-ability and TopTwoSpokesAngle in both experiments. By graphing the logistic regression between these variables, a lot of information can be used by the designer to improve the product. Figure 12 shows this logistic regression; the probability serial represents the probability of a good outcome at every value of the independent variable (the design parameter). The affordance response serial represents the categorized evaluation (user response transformed to a binary response) from users. This means that if an bending of well-nigh 150 degrees is chosen for the steering bicycle, then there is virtually an 80% chance that a user would rate it as a good design. A expert design, as mentioned before, represents qualities of 1 to 3 based on the 7-point Likert scale used in the experiment.

Effigy 12. Plough-power versus TopTwoSpokesAngle; experiment ii.

This can provide valuable information to a designer. Instead of trying to select one value for a design parameter, the designer can focus on a range of values that would make the users have a positive perception of specific affordances in the production. As mentioned earlier, the designer can use the optimized solutions obtained from experiments. However, these solutions may demand to exist changed to meet other design criteria. The relationships between affordances and design parameters can be used to change pattern parameter values and predict how users will perceive product affordances.

Too helping designers to target specific affordances with design parameter changes, these relationships give clues on how users make use of the product. The fact that turn-power is related to the angle between the top two spokes means that users use the spokes to plow the steering cycle. This data could be used to identify new affordances that could better the usability of the steering wheel. For case, designers could improve the grip on the spokes of the steering cycle to brand it easier for users to turn the wheel.

As tin exist seen in Table iv, design parameters tin exist related to multiple affordances.

Figure 13 shows the relationship between SeeThrough-power andTopTwoSpokesAngle for experiment 1 (similar results were obtained in experiment 2). The results of this human relationship make sense, as the larger this angle becomes, the easier it is to encounter through the steering cycle, making it easier to see the gauges on the dashboard in the cabin. HandRest-ability is also related to TopTwoSpokesAngle, which means that users remember of resting their easily on the spokes rather than on the ring of the steering wheel.

Figure 13. SeeThrough-ability/TopTwoSpokesAngle; experiment 1.

Figure 14. SeeThrough-ability/RingThickness; experiment 2.

Effigy 15. Protect-ability/RingThickness; experiment 2.

The thickness of the ring (RingThickness) is related to SeeThrough-ability and Protect-power (meet Figures 14 and xv). Unlike the relationships shown earlier, the SeeThrough-ability relationship is inversely proportional. This means that as the design parameter value increases, the probability of a good assessment decreases.

It should be noted that due to the fact that six users were employed in each experiment, the results practise not represent the perceptions of 'all' finish users, and therefore it is not expected that all experiments volition yield the same relationships every bit seen in Tables 4 and v. The goal of these experiments is not to characterize the perceptions of all individuals; for that to happen, larger crowds need to be used in each experiment (which is possible with ABIGA).

4.iv Design parameter values to target specific affordances

For the relationship shown in Figure 13, if the designer wants to know the angle between the top two spokes for which the probability of acceptance is 65%, Eq. (1) can be used to find information technology (the coefficients are given in Tabular array 6). This gives the designer a range of values for which the probability of acceptance is equal to or higher than 65%,

(ii) $$\begin{eqnarray}\ln \left(\frac{0.65}{1-0.65}\right)=-2.25213+0.0167285x,\terminate{eqnarray}$$

(iii) $$\begin{eqnarray}x=171.6\text{ degrees}.\end{eqnarray}$$

The result shows that if the designer chooses whatsoever value betwixt 171.6 and 250 degrees, most users would charge per unit the SeeThrough-power of the steering bike as skillful. This tin can be carried out for all of the relationships found through logistic regression analysis, giving designers a lot of freedom when choosing values for different design parameters while making sure that their decisions will be perceived as 'good solutions' by the users. It should be noted that the equation higher up tin just be used within the range of design parameters that was tested with users. Extrapolation cannot be performed with the logistic regression results.

4.v Multi-objective trade-off analysis using affordance/blueprint parameter relationships

The relationships in Figures 14 and 15 show a disharmonize between Protect-ability and SeeThrough-power with respect to the same design parameter (RingThickness). The Protect-ability and RingThickness relationship is positive. This means that if the RingThickness is increased, the perceived Protect-ability quality improves, but the perceived SeeThrough-ability worsens. It is the designer's choice to favor a detail affordance. This trade-off analysis can be made using the results from these 2 relationships.

For example, if the probability of acceptance is set at 0.vii, the RingThickness tin can be whatever value in the range (Reid et al. Reference Reid, Gonzalez and Papalambros2010; Nguyen et al. Reference Nguyen, Guarneri, Fadel and Mata2012), according to the SeeThrough-ability relationship. Nonetheless, co-ordinate to the Protect-ability relationship, the RingThickness value tin be any value in the range (Loh Reference Loh1996; Orsborn et al. Reference Orsborn, Cagan and Boatwright2009). The designer here tin clearly see the trade-off in choosing a ring thickness value. In this example, the designer can choose to favor 1 of the conflicting affordances.

4.6 Applications of ABIGA

As suggested throughout the paper, ABIGA can be used in the design procedure to help designers to optimize the shape of a product because the input from end users. The tool tin can as well be used in the redesign of products where the compages is already defined but the design parameter values tin be optimized. This newspaper focuses on the usability aspect of design, which is the reason why end users were selected as the type of user. Even so, the web application can be used on other types of users to optimize the design based on the input of that specific group. For example, manufacturing and assembling users will have their own set of affordances related to manufacturing and/or assembly. The needs of different types of users could potentially change the shape of the artifact.

4.7 Limitations of the current tool

The application optimizes the shape of the artifact. There are no changes in the topology. All the same, this would be possible if the designer had come upwards with different topologies (or configurations) and would like to know the optimal topology configuration. The features can be coded as existing (i) or not existing (0), effectively assuasive the GA to work with multiple solution configurations.

The set of affordances shown to the users is currently stock-still. The application can exist modified to let users to suggest affordances that might be added in the middle of the optimization run. This could aggrandize the solution space by calculation features to the products that fulfill the suggested affordances.

The tool currently does not allow for three-dimensional product representation to be shown to users. The use of iii-dimensional renderings of products could improve the ability of users to assess the quality of their affordances. This will be explored in future experiments.