1. Purpose
In the various manufacturing fields including the building field, the form design is devised by the designer. However, the designer's idea range depends on his personal background such as education or environment, so it has been pointed out that his design idea is controlled by his background. The facade of the building controls the visitor's impression, so the image of the design is very important. Various studies such as the constitutions or histories of the facade are performed, but the facade design support system has not been developed. Therefore, if there is an idea generation support system which is defeating a conventional idea of the designer and expresses the detail of the facade, it would be effective. The purpose of this study is to develop the idea generation support system giving “discovery” to the designer for the facade design of an office building. The purpose of the previous research was to verify the difference in the evaluation number of individuals of IDE. This study analyzes the form elements constituting a facade and proposes new algorithm IDES based on IDE, in order to create many various design forms efficiently. The multiplicity of design by IDES and the effectiveness of the system are verified in comparison with those by IDE.
2. Methods
58 form elements are extracted by collecting and classifying the image of an existing office building. 58 form elements are replaced to the individual's vector of IDE and IDES, and became incorporated into IDE and IDES algorithms. 3D images of facades by IDE and IDES algorithms are expressed by VRML. The verification of this system was performed by 20 students learning architecture.
3. Conclusions
In IDES, 80% of all creations were created as different designs without duplicate compared with the IDE. IDES is effective algorithm as design idea support system, because the “very good” design was created about 2 times compared with the IDE. The classification of form elements extracted in the previous research is proper, because the facade designs created by this system are realistic and diverse.

View full abstract

Purpose
The computer performance is improving, and the various system optimizations by using the evolutionary computing are developed in recent years. Though the forms of large span roofs can be built free, the mechanical performances of the roofs are greatly different according to those forms. Moreover, though the aesthetic sensibility is demanded, the studies of morphogenesis which consider the aesthetic sensibility and mechanical rationality are little. We developed an idea support morphogenesis system that satisfies the designer's image and proposes the form with a high mechanical rationality, by using the Genetic Algorithm. However, the form that can be proposed by the existing research is limited in a symmetric form, and the proposed form does not have a guarantee that satisfies the designer's image. To solve these problems, and to develop a more effective system, the improvements of this study are as follows;
1) To enable the input of asymmetric form image
2) To propose Pareto solutions as two purpose optimization problem of “Image” and “Equivalent stress”, and to develop the system to which the designer can decide the final form.
3) To obtain the design knowledge concerning a morphological change necessary to improve the mechanical rationality.
Methods
The image of an old system was input by two directions (the direction of X and the direction of Y). In the new system, the complex forms are created by the image input for three directions (direction of X1, direction of Y1, and direction of Y2) or four directions (direction of X1, direction of X2, direction of Y1, and direction of Y2). The form is defined by control points of Spline curve. The form created by the image input is changed slightly by moving control points. This system creates various forms by changing the moving quantities of control points and proposes Pareto solutions of “image” and “mechanical rationality”. From the execution results, we analyze how the form is changing by differences of the moving quantities of control points.
Conclusions
By changing the control point magnification large, the equivalent stress became small though the image difference from an initial form became large. “Image” and “Equivalent stress” are in the relation of the trade-off. It became possible for the designer to select the form which satisfies both his image and mechanical rationality by proposing Pareto solutions. From an existing research, the design knowledge has obtained, that is, a mechanical rationality is improved when the curvature of form becomes a single curvature. From the execution result of model 1 and model 2, we could reconfirm this knowledge.

View full abstract

This research proposes how to divide or distinguish a curved surface into gradual domains on the basis of the points arranged on the curved surface. It is shown that this method can make various forms from the form constituted in line to the form of field composition based on the free form curved surface. The fundamental idea is as follows. First, the creation method of Voronoi diagram is applied by making the points on a curved surface into generators. Next, space is divided by a tetragonal lattice and made discrete. And the random graph limited to the tetragonal lattice is created, and the distance on a curved surface is measured using the graph. The form from line composition to field composition is created by creating the Voronoi diagram which had the ambiguous boundary using the above method. And the examples applied to waving curved surface, rectangular prism, hemisphere and free form curved surface were shown. In recent years, the construction of free form architecture is becoming realizable by progress of technology, and this method will be useful for development of more free form.

View full abstract

With diffusion and development of information-related technologies in recent years, geographic space information services, typified by car navigation, are widespread. However, provision of map information, which is the basis of geographic space information services, depends on on-site investigations conducted manually; therefore, there is a problem that the map information cannot catch up with the changes in the real-space information, such as the opening of a new store. Therefore, the authors have been conducting a study on text mining to automatically collect store information such as store names, addresses, and the industry types from the natural language, with focus on the information about the opening of new stores on the Web. However, when conducting the text mining using only the keywords related to new store opening, we were faced with two problems. One is that the differences in the word appearance tendency by industry type or the combination of words before and after one another become a noise, which causes the accuracy of the collected store information to fall. Another problem is that since the precision of the information sources that may contain typing errors or false information is not guaranteed, it is difficult to judge whether the data is correct or not if there is inconsistency in the contents. This study aims to overcome these problems by mining using the keywords, and by considering the fact that there is a strong correlation between the industry type of a store and the geographic characteristics of the store location. By experimenting, we verified that this study considering the geographic characteristics of the store location was effective to collect the new store information.

View full abstract

It should be an issue that the central area in local city declines commercially because of the extension of city area. Therefore, the urban development policy is proposed to activate the central area of the city. The urban development policy is not often discussed on the basis of numerical value of evaluation in the real world. The ordinary aggregate model cannot describe propalg the complex interaction between urban activities. The complex system modeling approach is introduced in the study. The artificial society model is created with citizen agents and commercial retailers. In particular, the decision of citizen agents for shopping and leisure activity place and travel mode can be formulated by fuzzy reasoning. On the other hand, the commercial attraction degree can be influenced by the commercial sales and number of visitors. The decrease of activities in the central area without urban development policy can be observed in the artificial society. On the contrary, the developments for several different activities in the central area can be found with introduction of fundamental urban development policies. Furthermore, it is realized that the comprehensive urban development policy with combination of policies should be superior to the single policy introduction. Finally, it is concluded that self-organized development process of local city can be realized from the accumulated activities of individual urban citizens with referring to the observation in the artificial society model to evaluate urban development policy.

View full abstract

In this paper, supporting system for regeneration of existing school buildings is proposed. Many existing school buildings have problems of seismic performance and function. To solve these problems at the same time, it proposes the design supporting system with Multi-Objective Genetic Algorithm. The conditions to solve these problems are presented as follows: (i) the number of student remains, (ii) making extensions to a building for setting up multipurpose rooms, (iii) conversion of school buildings. The problems of seismic performance and function for regeneration are formulated respectively for these models. Finally, it aims to present the solution that can be designed and suit the regeneration demand of the existing school buildings.

View full abstract

Recently, in Japan, utilizing methods on the stock of dwelling houses become an important issue to be solved in the society where population is shrinking. In order to resolve the problem, “Reduction Method” is considered to be one of the effective methods. Reduction method is used in order to improve dwelling environment by reducing building volume. In case of location problems of dwelling houses, researches on evolutionary computation methods are studied and its effectiveness for obtaining an optimal solution from enormous solution space in a short time is also clarified. In this paper, simulation systems of reduction method on dwelling houses are developed by using Classifier System: a reinforcement learning system. By using a classifier system, layout plans of dwelling houses are obtained as results of simulation, and rules of reduction method are obtained, therefore, processes of reduction method can be observed. Furthermore, a reduction simulation is also performed using Genetic Algorithm and the layout plans of dwelling houses created by the rules of reduction method are verified and discussed by comparing with the plans obtained as a result of Genetic Algorithm simulation.

View full abstract

In recent years, legion of construction waste and paradigm shift from the mass production are recognized as issues of the residential housing production, and thus provision of housings that meet the diverse demands of the residents are in need. Concerning the issues, an example is the “200 year residence” leaded by the government. The policy endorses the high durability of housings, which further curb the construction wastes; however, it is difficult to cope with the long term transition of living spaces. Unit residence which aims to reduce costs through reutilization of structural materials and cope with the change in lifestyles exists, but the circulation is done through suppositions of future usage, which later causes restrictions in building programs. Therefore, we propose a “20 year residence” which diverts and circulates structural materials while meeting the changes in demands of the residents. The research aims to promote utilization and improve circulation efficiency of structural materials, therefore formularization of multi-objective optimization are conducted by treating the total number and the ratio of structural materials quantitatively. Based on the solution derived from the multi-objective optimization, building framework is configured to discuss the possibility of residential housing framework design different from the past. Furthermore, through numeric simulation supposing the diversion and circulation of structural materials between residential housings, the usability of the “20 year residence” proposed in this paper.

View full abstract

Recently, the bridge management is required for keeping the safety of daily life. On the other hand, it is difficult to sustain the management in the future due to the short of engineers and the budget cut. The long term planning for management is attempted to reduce the cost for maintenance and improve the safety. However, this planning has some problems due to uncertainties and the equalization of cost. The uncertainty in the prediction of degradation may cause the change of plan. The purpose of this paper is to develop the decision support system for long term planning. In the long term planning, it is undesirable to increase the maintenance cost due to the change of plan. Therefore, in this paper, an attempt is made to propose the dealing method against uncertainties. The proposed method estimates the flexible term for maintenance year without increasing the cost. In the proposed method, a plan considering the preventive maintenance is searched by Genetic Algorithm (GA). In this step, the obtained plan is expected to be able to change the maintenance year without increasing the cost because the search of GA has the fuzziness in a problem which has various suboptimal solutions. The estimation of the flexible term for the obtained plan is attempted by applying the numerical simulation. Therefore, the plan obtained by the proposed method is expected to improve the safety against uncertainties without increase the cost by following the flexible term. Numerical examples are presented to demonstrate the applicability of the proposed method.

View full abstract

There is a possibility that multiple lifelines get damaged at the same time when a large scaled disaster happens. This causes much damage to our lives, therefore, we urgently need a plan to recover from the disaster. However, since each lifeline has close relationships with the other lifelines in terms of their functions - there are several constraints among them in the process of restoration. Therefore, it is difficult to promptly make a proper schedule for disaster restoration. With this background, in this study, we aim to efficiently make a disaster restoration schedule taking account of several mutual relationships among lifelines by means of genetic algorithm. In using genetic algorithm, we propose a gene expression taking account of squad allocation and the priority order of each lifeline's restoration in its coding, and then use the gene model in the calculation process of making a schedule. In addition, we have introduced a local search, i.e., random flip method, into the algorithm for an efficient schedule within a short time.

View full abstract

Advanced traffic controller for urban expressway would be recommended corresponding to the recent progress of information technologies as well as soft-computing. On the other hand, it has summarized as significant results of researches that the essential traffic control for urban expressway can be discussed based on the fuzzy control formulations. The flexible and robust traffic controller can be created with the formulation of fuzzy logic. The optimization of parameters for efficient traffic controller can be performed by genetic algorithm as well. The fuzzy traffic controller is evaluated through the estimation of practical condition with a traffic accident by the traffic simulation for urban expressway in the study. The rules for the fuzzy control can be updated with the knowledge about traffic control for accident, as the structure of fuzzy control is flexible. The fuzzy controller for traffic accident condition is applied and evaluated with the traffic simulation. The advanced inflow controller provides the efficient traffic on the urban expressway even with the complexity even in the real world. Finally, it would be confirmed that the fuzzy control with knowledge acquisition is effective as traffic control for urban expressway.

View full abstract

In this research, we create the lighting system that users can order brightness sensuously such as “very lighten” or “little darken” by using Actor-Critic algorithm. This system needs to learn the sensory scale so that it can act depending on 2 kinds of statuses, which are the demand of a user and the brightness around the user. We built up Actor-Critic algorithm with 2 kinds of Actors applying to 2 kinds of states. In this paper, the overview and effectivity of this algorithm are shown.

View full abstract

The traffic flow information of humans and goods is widely put to use as infrastructural information in urban planning such as traffic plans or area marketing. However, traffic flow information depends on the questionnaire survey of the public transportation users or the visual inspection by the inspector on a street corner, causing a problem of costing a lot of money. Therefore, studies have been conducted to automatically measure traffic flow information using a laser beam sensor or a digital video camera. The former method using the laser beam sensor makes it possible to perform a traffic survey of a movable body with high precision in any environment. However, it is impossible to obtain information of its appearance and thus grasp correct attribute information of the movable body itself. On the other hand, with the latter method using the digital video camera, information about the appearance of the movable body can be obtained and analyzed. However, there is a problem that the measurement precision goes down when the photographic environment changes such as in lighting and sunshine, or under such environment as occlusion occurs due to the overlapping of movable bodies. Thus, this study realizes accurate measurement and acquisition of human flow information with attribute information even under difficult photographic environment or where occlusion occurs, by focusing only on human figures and using a range image sensor to which the characteristics of the laser beam sensor and digital video camera are added.

View full abstract

In recent years, with the spread of digital cameras and development of the image processing technology, it has become possible to take a panoramic photograph easely. There is a rapidly growing need for panoramic images, for example, for recording the progress or products of construction works, because the panoramic image makes it possible to represent a wide range of scenery in one picture. However, it is unfavorable if the panoramic image contains a heavy machine or workers having nothing to do with the structure during the construction or at the completion in carrying out precise progress control or minute quality control. This study proposes a method to automatically generate a plane panoramic image by finding out the optimum image for the background out of plural digital images in time-series, and removing movable objects in the end, using the time series characteristics of the digital moving image shot freehand with a digital video camera. First, alignment is performed for the frame image obtained from a digital moving image. Next, STV (Spatio Temporal Volume) is generated from the frame image after calibration. Finally, a plane panoramic image is generated, from which the movable objects are removed, by estimating the picture elements in the background from EPI (Epipolar Plane Image) of STV. Finally, the validity of the proposed method and the usability of the plane panoramic images are verified.

View full abstract

In recent years, with the spread of digital cameras and development of the image processing technology, it has become possible to take a panoramic photograph with ease. Especially, since a 360-degree panorama image has eminent qualities in representing periphery environments from the camera point of view, it has a high affinity in urban town guide systems as well 3D demonstrations. With the advent of panorama viewing systems such as Google Street View, the sphere of utilization of this technology has gone dramatically expanded. Therefore, the using range dramatically expands with spreading panorama view system like Google Street View. However, it is unnecessary information to grasp the surrounding situations if the panoramic image for services used by many people contains people or vehicles having nothing to do with, and unfavorable to protect privacy. Therefore, this study proposes a method of generating a 360-degree (whole circumference) panoramic image complemented with the movable objects in the background, by identifying the area of any movable objects from the digital photographic image shot freehand with a digital camera. First, characteristic points are generated on each frame of the moving image. Next, using these characteristic points, the corresponding points between the images are obtained and aligned with each other so that they overlap with the base photographic image. Then, the movable object area is identified from the corrected image sequence. Finally, a 360-degree panoramic image with the movable object area complemented in the background is generated by projecting the image on the cylinder surface while complementing the movable object area of the image from which the movable objects are extracted. Thus the validity of the proposed method and the usability of 360-degree panoramic images are verified.

View full abstract

While construction CALS/EC is promoted, it is important to make full use of the drawings over the life cycle of a public works project to improve its productivity. However, drawing formats are different from phase to phase in the whole life cycle. For a road project of Ministry of Land, Infrastructure, Transport and Tourism (MLIT), for example, Survey is managed with expanded DM data while design and construction works are managed with CAD (SXF) data, and maintenance with GIS data. Therefore, MLIT developed a system for converting expanded DM to CAD/SXF, and CAD/SXF to GIS, to make full use of the data. Regarding the public works projects of the local governments, on the other hand, formats for survey, design, and construction works are similar to those in MLIT. However, maintenance is often managed with expanded DM. Therefore, in the local governments, there is a need for converting CAD/SXF of the construction works products into expanded DM so that the expanded DM to be used for maintenance can be efficiently provided. However, in many foreign countries and Japan, standard specifications or conversion systems for converting CAD/SXF to expanded DM are not provided. In this study, we developed conversion specifications from CAD/SXF to expanded DM, and a system for converting CAD/SXF into expanded DM. In addition, we prove the usability of the method devised in this study by converting CAD/SXF into expanded DM using the drawings used in public works projects.

View full abstract

Construction CALS/EC is introduced in public works projects over the whole life cycle for the purpose of reduction of the construction period, quality guarantee, and cost reduction. Especially recently, environmental improvement of the information-oriented construction using 3D data attracts attention for engineering works stage. In the construction site, it is expected to create 3D model from the point cloud data obtained by surveying the site with the total station or the laser scanner. However, it is necessary to solve the problem of how to handle the great amount of point cloud data obtained at the survey based on their characteristics, as well as the problem of required accuracy for the information-oriented construction. The authors have devised a technique to make 3D model of the river embankment, which satisfy the required accuracy of the information-oriented construction from a large quantity of point cloud data. This technique used the point cloud data of the river embankment measured with the laser scanner and DM data of the river embankment. Then, the boundary line (break line) between the crown surface and the slope face of the river embankment is extracted automatically to create 3D model on CAD. However, the following problems were revealed; the freshness depending on the update cycle of DM data, and wrong extraction of a break line due to the noise such as the vegetation between the crown surface and the slope face of the river embankment. In this study, we devised a method for automatically extracting a break line by inferring the crown surface area of the river embankment from the point cloud data, and a method for removing any point cloud data that are the noise such as the vegetation on the crown surface of the embankment. Then we performed evaluation experiments and proved the usability of the devised solution.

View full abstract

The famous Greco theorem (G.H. Greco, Rend. Sem. Mat. Univ. Padova, 66, 21-42, 1982) and our new method of proof lead us to an improvement and another proof of Sugeno-Narukawa-Murofushi's Riesz type integral representation theorem (M. Sugeno et al., Fuzzy Sets and Systems, 99, 205-211, 1998) for comonotonically additive functionals.

View full abstract

Two types of vector-valued Choquet integral models are proposed. Vector-valued Choquet integral models are vector valued functions calculated by m times Choquet integral calculations with respect to the m-th fuzzy measure vector. This model is an extension of the product of a matrix and a vector. Logical vector-valued Choquet integral models are extend to functions for which the input and output vectors are vectors with coefficients in the interval [0, 1] and fuzzy measures are set functions that map the interval [0, 1]. If the sum values of the set function values are equal to 1 for all subsets of the domain of the fuzzy measures, then the sum of the output values is 1. To introduce the symmetric difference expressions, some non-monotone fuzzy measures can be transformed to monotone fuzzy measures and can be interpreted on the basis of Shapley values etc.

View full abstract

We propose a novel quantization method for human likelihood or preference using several sorting results. We estimate an evaluation function by assuming that the data is sorted on the basis of observed function values and that the sorting order varies from person to person according to the observation error for the function. We present tests to evaluate the proposed method by using random data and sorting data for gray images. These will show that our method is valid enough when the data have certain variety.

View full abstract