This paper presents a topology optimization method for the design of elastic metamaterial based on a high-frequency homogenization method for elastodynamic equation. First, we briefly explain the high-frequency homogenization method for elastodynamic equation to estimate the macroscopic properties and dispersion relation of elastic metamaterial. Next, optimization problem is formulated so that optimized unit cell design of metamaterial exhibits hyperbolic dispersion relation. Two components of second-order tensor in the homogenized equation, which represents approximated dispersion relation of elastic metamaterial, are used in the objective and constraint functions. We use level set-based topology optimization method to solve the optimization problem. We offer a two-dimensional numerical example to show the validity of our method. Optimal configuration obtained by the proposal method consists of steel, and it exhibits hyperbolic dispersion around the point corresponding to the target eigenmode in wavevector space due to a local resonance induced in the unit cell of elastic metamaterial.

In the indoor space that a person lives or works, a ventilation function is demanded from the perspective of thermal environment. However, the noise from the external environment becomes the problem by establishing the through-holes such as windows on the wall to have a ventilation function. In late years studies of acoustic metasurface which is controllable by the sound wave such as soundproofing and the acoustic absorption function are accomplished. A ventilation function and a soundproofing function were contradicting functions, but a study of acoustic metasurface which balanced a soundproofing function with a ventilation function has been accomplished. Based on the study, We designed acoustic metasurface of the wall having the through-hole which enhanced a soundproofing function and a ventilation function by an Topological optimization. This paper describes the shape of the acoustic metasurface and the inspection result of an experimental noise reduction effect.

This study presents a topology optimisation for far-field characteristics of elastic waves scattered from periodic structures. We first show that the far-field characteristics can be expressed with a boundary integral on scatteres by exploiting periodic Green's function. Then, a topological derivative of this integral is derived. We finally design a wave motion converter by using the topological derivative and a level-set-based topology optimisation.

We propose a topology optimisation for designing acoustic diodes which realise uni-directional transmission of sound wave. Several types of acoustic diodes based on different physical mechanisms have been proposed so far. In this study, we focus on a phononic-crystal-based acoustic diode. We can design such an acoustic diode by maximising so-called diodicity i.e. the difference between transmittance of a wave from one direction and that from the opposite direction. We derive the topological derivative of the diodicity and discuss a fast boundary element method to evaluate the topological derivative numerically. With the topological derivative, optimal design of acoustic diode is explored by a topology optimisation with level-sets of B-spline surfaces. We present numerical examples showing that the proposed topology optimisation can find a periodic structure with high diodicity.

This paper describes a level set-based topology optimization method for unsteady state Navier-Stokes flow. Many methods for topology optimization applying to fluid mechanics have focused on steady state while most of all fluid phenomena should be considered as unsteady state problem. Therefore, fluid flow problem for topology optimization method should also be solved as unsteady state problem. In addition, wall treatment of fluid field such as velocity has to be considered as another theory such as boundary layer problem. In order to consider wall treatment, applying an immersed boundary method to boundary condition on wall is an effective method so that we apply the immersed boundary method on wall of optimized structure in our method. As a result, we can obtain the optimized structure by applying our method. Some examples are provided and the effectiveness of the method is discussed.

Structural optimization has been successfully used in many industries. In particular, topology optimization, which has the most potential for exploring optimized structures, can be applied to a wide range of structural optimization problems. However, most existing topology optimization methods for fluid flows problems are based on the finite element method or the finite volume method, and have difficulty in handling problems which include free surface. Moving particle semi-implicit (MPS) method is one of the particle methods, which can be used to analyze incompressible free surface flow without surface tracking. This Lagrangian mesh-less method, proposed by Koshizuka, et al has been proven to be useful in a wide range of engineering applications, such as numerical analysis of turbines and mixers. This paper proposes a topology optimization method using MPS method for fluid dynamics problems. Several numerical examples are provided to demonstrate the validity of the proposed method.

In order to obtain a global topology optimal solution at supersonic flow conditions, it is necessary to solve efficiently high dimensional optimization problems with nonlinear discontinuous performance functions. As a countermeasure to solve these complicated optimization problems, we developed a novel optimization method which uses machine learning techniques such as DBSCAN clustering method and Support Vector Machine two-classes classification method. The developed method can explore the global optimum solution without probabilistic information such as variance calculated by Gauss Process Regression. In the test cases of multimodal/discontinuous functions, the developed method could obtain global optimal solutions superior to a conventional optimization method utilizing the EI acquisition function. Since the developed method uses only the predicted performance values to explore the global optimal solution, it can combine with any response surface methods such as Kernel Ridge Regression, Neural Network, Support Vector Regression, RBF Network, Reverse Distance Weighting Method and so on. In other words, when the developed method combines with response surface methods with higher adaptability to high dimensional problems, it will be able to solve high dimensional optimization problems efficiently.

This paper presents a design approach for fin-and-tube heat exchangers, which deal with complex physical phenomena. In this study, the shape, arrangement, and number of the protuberances on the fin are designed from topology optimization using a simplified model. Although the simplified model incorporates the key factors of the heat exchange performance, the physical phenomena in the simplified model cannot be completely matched with that in the high-fidelity model. Therefore, a variety of design candidates are generated by the optimization on various parameters, and then the satisfactory solution is led from these candidates by the high-fidelity analysis. A manufacturable fin is designed from the satisfactory solution, and it shows better heat exchange performance compared to reference fins.

In the topology optimization of fluid problem, there are many prior studies targeting laminar flow, but due to the complexity of the discussion and the high calculation cost, there is little preceding research on turbulent flow, which accounts for most of the real flow. Therefore, as a fundamental study on topology optimization for turbulent flows, this paper presents the verification of Brinkman model, which is generally used to topology optimization for laminar flow to simplify treatment of walls of flow path. In this study, several flow tubes that are typical examples of structural optimization problems are adopted to compare Brinkman model’s flow and actual turbulent flow. Since it is difficult to describe and analyze turbulence using Navier-Stokes equations, low Reynolds number k-ε model which is a kind of RANS equations is used. The finite element method is applied to solve that governing equations.

In the topology optimization problem, quantities defining design regions or boundary conditions are treated as design parameters. However, these design parameters can be variable from the viewpoint of application to mechanical design. The value of the design parameters may have a significant influence on the performance of structure. In this research, we proposed an approach for finding appropriate values of design parameters in the topology optimization problem. In the proposed method, the problem of optimizing the structure and design parameters is replaced with a set of two problems, which avoids a difficulty in topology optimization problem due to the change of the design parameter. By introducing metamodeling, design parameters can be optimized with low computational cost. We apply the proposed method to PMSM design problem and show effectiveness of changing design parameters.

A new structural optimisation method using shape and topological derivatives simultaneously is proposed. Configuration of the design object is expressed through the level set method in which structural boundary is expressed as a zero-value isosurface of a level set function. With the help of the level set method, configuration change of the design object including creation of new holes is associated with the perturbation of the level set function. By using such an association, shape and topological derivatives can be naturally combined. Through a numerical example of minimum compliance problem, the effectiveness of the method is confirmed.

In product design, conceptual design is important step that greatly affects functions and shapes, and it determines the direction of the products. In addition, intention of designer is reflected the most in the process of conceptual design. Therefore, the value of a product is increased by having the intention of designer, for instance, uniqueness, preference, playing, and aesthetic sense as technology and design. It is summarized as using Kansei value in “Kansei Value Creation Initiative” published by The Ministry of Economy, Trade and Industry and it is necessary to visualize Kansei value as a requirement for the concept design stage. We acquire EEG by using BMI from Frontal Cortex expressing sensitivity and Supplementary Motor Area performing physics simulation. Furthermore, we create shape by FSD to design based on Mechanical Kansei. By combining these two, it is possible to obtain a layout using sensibility. The shape and layout imagery creation system is discussed with the verification result and the creation result.

Microwave resonators are devices used for electron spin resonance spectroscopy which is expected to be one of the essential ingredients for future quantum technologies. The performance of the microwave resonators depends on the magnetic flux density norm in the sample space of the devices. Conventional microwave resonators have been designed to increase the magnetic flux density norm by resonance in the sample space. However, since the structures have been made by trial and error, there would be further improvement in performance of the devices. Thus, in this research, microwave resonators are designed by level set-based topology optimization. First, the design model and the governing equations of electromagnetic wave propagation problem are explained. Second, we propose a topology optimization method for the design of microwave resonators where two objective functions are defined. Third, the optimization problem is formulated to avoid local solutions because of frequency-dependent property. Finally, a numerical example is provided to confirm the usability and validity of the proposed method.

The purpose of this study is to establish the designing scheme of simplified structure of superconducting coils for particle beam accelerators, which can be easily manufactured, by using of the procedure of topology optimization. We represented the coil configuration with the density function corresponding to the conduction cable density per cross-sectional area. We set three axes of design requirements for particle accelerators; magnetic field homogeneity, magnetic field strength and small cross-sectional area of the coil. We formulated the objective functional and developed the optimization algorithm to solve the multi-objective problem. In the algorithm, we applied the Heaviside-projection method after PDE-based filtering to avoid the gray scales in the density function and excessively complex structures. Finally, the validity of the proposed scheme was confirmed with numerical examples of coil designs.

Although the rare earth material is actively applied to the electric vehicle, train, and so on, the cost of the rare earth material rises more and more. Therefore, the synchronous reluctance motor (SynRM) which does not require permanent magnet is remarkable from the viewpoint of lower price and high efficiency. However, the torque characteristic of SynRM is much lower than that of interior permanent magnet synchronous reluctance motor (IPMSM). Then, permanent magnet assisted synchronous reluctance motor (PMASynRM) which bonded magnet was attached to the rotor structure of SynRM has been proposed. Because bonded magnet is applied to PMASynRM, the manufacturing cost of PMASynRM is lower than that of IPMSM. In this paper, the structure of PMASynRM has been drastically changed by topology optimization. As the result, improvement of torque characteristic is achieved in the optimized structure.

Power electronics equipment is used for electric power conversion in hybrid cars, which are attracting attention in recent years due to high environmental performance. Since the electromagnetic noise emitted by the power electronics equipment adversely affects other electronic equipments and the control circuits of power electronics equipment itself, an electromagnetic compatibility (EMC) standard is defined in order to prevent this harmful effect. Normally, a noise filter circuit for suppressing a high-frequency noise current is provided around the power electronics device in order to achieve a design conforming to the EMC standard. Since the performance of the noise filter is influenced by the magnetic flux generated from the current loop on the circuit, the design of the conductor pattern shape is important for improving the performance of the circuit. However, it is difficult to design circuits by quantitatively predicting the influence of magnetic flux. To solve such problem, filter performance can be improved by using topology optimization that enables structure optimization with high degree of freedom. In addition, many electronic circuits are designed as double-sided boards with different patterns on the front and back sides, and the design considering the influence of the magnetic flux becomes more complicated. In this research, we aim to obtain higher noise suppression performance by simultaneously optimizing the front and back of the filter substrate as the design domain by creating the boundary matching mesh on both sides.

In this research, we study on updating method for the orientation of anisotropic material in topology optimization. In the topology optimization of anisotropic materials, the material distribution and the orientation of the fiber are treated as design variables. In this research, we focus on the optimization of the orientation. First, we express the orientation as a vector field, and define elastic coefficient tensor of an anisotropic material. In the previous research, the upper limit was set for the nom of the orientation vector field, which results in restriction on updating the orientation. Thus, there is a possibility of falling into a local solution. Therefore, in this research, we formulate the optimization problem with constraint on the orientation vector field based on the augmented Lagrangian method. Finally, we present a numerical example for the stiffness maximization problem of the linear anisotropic elastic material, and confirm the validity of the proposed method.

This paper presents a method to derive the design sensitivity by the adjoint variable method based on the lattice Boltzmann method (LBM) in order to perform a topology optimization using LBM for the structural mechanics problem. Since LBM is described as an algorithm based on a simple full explicit method, it has the feature that it is easy to introduce it to large-scale parallel computation. Various models such as fluid model and thermal model using LBM have been proposed so far. However, there are few cases where elastic bodies have been analyzed using LBM, and there is no example that a topology optimization was done. In this paper, for elastic problems, we construct an optimization formula based on the lattice Boltzmann equation and the adjoint lattice Boltzmann equation.

近年、ナノ構造に超短パルスを照射した際の時空間ダイナミクスが注目を集めている。本研究では、所望の超短パルスの時空間プロファイルを得るための設計手法として、時間領域のトポロジー最適化手法を提案する。超短パルスを直接取り扱うため、時間領域のマクスウェル方程式を用いて最適化問題の定式化を行った。目的関数は超短パルスの包絡線を用いて定義した。最適構造の境界を明確に表現可能なレベルセット法に基づいてトポロジー最適化を行った。数値例として、パルス幅を保持したまま入射パルスを集光するナノ構造の最適化設計を行い、本手法の有効性・妥当性を確認した。

In this talk, we consider stiffness maximization problems where the effect of body force is dominant. In these problems, one needs to choose the objective function carefully in a manner that it represents the desired stiffness appropriately. For example, “all void” solution is the trivial optimal solution of compliance minimization problem where only the body force is loaded. This may cause numerical difficulty and the optimal result may not be a desirable one in applications. We thus focus on the stiffness maximization problems for elastostatic problems in 3D domains to consider the settings of the objective function when the body force is dominant. Specifically, we consider the compliance in whole domain and total displacement on a certain part of the boundary as the objective functions to observe the optimal shape and convergence history numerically. We show some numerical examples obtained with level-set based topology optimization method proposed by Yamada et al.

In this research, we developed a pattern-diagram optimization algorithm considering overtaking by express trains in suburban railroads. We first developed a traveling simulation algorithm that takes into consideration various constraints and existence of express trains, and optimized pattern-diagrams in suburban railroad by combining the constructed algorithm and genetic algorithm.

Due to maturation of science and technology, it becomes increasingly difficult to differentiate products in terms of performance, functional feature or price. Therefore, companies are required to differentiate their products in terms of subjective and abstract qualities such as aesthetic and ergonomics whose evaluation depends on customer feeling i.e. kansei. For years, various aesthetic design methods that consider customer kansei have been developed, but diversity of customers’ kansei is still a big problem. More specifically, since customers’ kansei and preferences are quite diverse, it is quite difficult to design mass-produced products that satisfy all customers. To overcome such problem, a new aesthetic design method considering variation of customers’ kansei is proposed in this report. In the proposed method, customers evaluate existing products using SD method and the relationships between their evaluation results and design parameters of evaluated products are approximated by response surfaces. Since impressions taken from a certain product can be estimated by using a response surface, optimal values of design parameters that give particular impressions to all customers which a designer intends are then explored. However, since different response surfaces are obtained for each customer due to the diversity of their kansei, satisfactory degree that represents the difference from targeted impressions is calculated for each customer and a multi objective design optimization considering the mean and variance of customers’ satisfactory degree is performed. In the case study, to evaluate the impact of variation of customers’ kansei on optimization results, the proposed method is applied to a simple mathematical model.

部材断面積を連続変数とするトラスの最適化の解は，一般に，多くの部材が異なる断面積をもつため必ずしも実用的であるとは言えない．本稿では，断面積が一様という制約を導入したトラスのコンプライアンス最小化問題に対し，大規模な問題例にも適用可能な発見的解法として，交互方向乗数法に基づく手法を提案する．

The gas in microfluidic devices is likely to be rarefied because the characteristic length of the channel is relatively short compared with its mean free path. Recently, some devices utilize the properties of rarefied gas flows in a positive way. Therefore, in this study, we propose the topology optimization for designing such devices. We use Bhatnagar-Gross-Krook (BGK) equation as the governing equation to analyze rarefied gas flows. Moreover, for the implementation of the boundary condition on the gas-sollid interfaces in the optimization process, the BGK equation is extended to the entire design domain composed of both the gas and solid domains, using the shape representation by the level-set function. We use the Hamilton-Jacobi equation and shape sensitivity to update the channel structure, according to the procedure of a level set-based topology optimization method. The shape sensitivity is derived based on the adjoint variable method. Finally, we show the validity of the method through a numerical example of pump design problem.

The draft tube of a hydro turbine has a role of discharging water to lower lake, and the turbine performance significantly changes according to its profile. Normally, the profile of a draft tube is optimized by CFD under constraint conditions at each power plant location. In this study, we attempted to replace the time consuming CFD process with response surface. The response surface method was RBF(Radial basis function). In order to verify the accuracy of the surrogate model, CFD was carried out for the optimized shape by the model. As a result, it was confirmed that both results of the model and CFD were in good agreement.

In this study, a solution to the shape design optimization problem of a 3D structure vibrating with initial stress is proposed, where a set of specified vibration eigenvalues with weighting coefficients is introduced as the objective functional. The eigen-equation with geometric stiffness term and the volume are considered as the constraint functionals. The problem is formulated as a distributed-parameter shape optimization problem, and the sensitivity function with respect to the shape variation is theoretically derived using the Lagrange multiplier method, the material derivative method and the adjoint variable method. The optimal shape variation is determined by the H¹ gradient method, where the shape sensitivity function is applied as a distributed force to vary the shape. The repeated eigenvalue problem is also considered by switching the original objective functional to the repeated one. With the proposed method, the vibration eigenvalue can be controlled to the target without parameterization of the design variables, while maintaining the smoothness of the boundary shape.

Heat conduction in nanostructures is ballistic process, whereas heat conduction at macroscale is diffusive. When a system has two types of materials, temperature discontinuities occur on the material interfaces. Thermal designs for electric devices in nanostructures utilize these unique phenomena, especially temperature jumps on material interfaces and they let us obtain innovative material properties. However, almost all conventional designs for nanoscale heat conduction problems have been dependent on heuristic and experimental approaches. In order to develop high-performance electric devices, we propose a level set-based shape optimization method for temperature minimization problems in nanostructures. First, we define an optimization problem for a nanoscale heat conduction problem based on the Boltzmann transport equation, then obtain a shape sensitivity through an adjoint method, which can precisely consider the temperature discontinuities on material interfaces. Next, we expand a level set-based shape optimization method to the design problem. Finally, we confirm the validity of our proposed method through a numerical example.

A variable fidelity concept is introduced in the proper orthogonal decomposition (POD) based re-parameterization approach to efficiently solve multi-objective aerodynamic shape optimizations. The re-parameterization approach enables to extract dominant shape deformation modes from a database of good airfoil designs and to reduce the number of design variables. Our proposed variable fidelity approach is realized by utilizing low-fidelity functional evaluations to select the good airfoil designs. The validity of the proposed approach is investigated in a multi-objective aerodynamic shape optimization problem of 2D airfoil.

In order to enhance the cooling performance of a vehicle, a front-end cooling structure is optimized with 3D-CFD and optimization algorithm. In calculation, the best design was explored and the airflow velocity passing through the heat exchangers increased by 12% in comparison with the baseline. The verification test is carried out with the prototype front bumper based on the calculated best design. The results showed the shape optimized bumper could be effective to improve the cooling performance.

The shock response caused by transient vibration often becomes a problem in various vehicles. In this study, the countermeasure for shock response of thin plate structure was designed by optimized calculation. The optimized calculation was conducted by topology optimization and size optimization. The topology optimization was conducted based on frequency response analysis. The result indicated that the maximum value of shock response spectrum (SRS) reduced by 73.5% compared with typical double cross rib configuration. The size optimization was conducted for abstracted model from result of topology optimization based on transient analysis. Finally, the optimized rib structure showed that the maximum value of SRS reduced by 52.7%.

This paper presents a formulation and solution of an identification problem of the muscle activities in an organ when a deformation of the organ was observed on boundary. This study is motivated by a medical interest of wishing to know the mechanism of swallow motion and cause of aspiration. In order to elucidate the mechanism, it is necessary to identify the contraction movement of the muscles. In this study, we assume that an organ consists of hyper elastic body, and the contraction movement is a finite deformation due to the generation of the anisotropic and inelastic strain. The strain is modeled as a function of a design variable which controls the contraction rate. Based on the assumption, we formulate a state determination problem of hyper elastic deformation by the compulsory displacement on the observed boundary using a given data. Using its solution, we define an objective function by the integral of the squared norm of the reaction force over the observed boundary, and construct an optimization problem seeking the design variable that minimize the objective function. Solution of the problem is presented as an iterative scheme using the H¹ gradient method for topology optimization problem of density type. Numerical examples for a cubic and cylinder bodies demonstrste that the deformations are reproduced by the identified design variables.

Because the performance of permanent magnet synchronous motor (PMSM) is strongly dependent on the quality of magnetization in permanent magnet, the operation for magnetization has to be accurately carried out on permanent magnet. However, the magnetization fault regretfully will occurs more or less. To decrease number of inferior permanent magnet with fault magnetization, identifying the distribution of magnetization in advance is considerably important from the viewpoint of high-quality-manufacturing of PMSM.

One of the methods to identify the magnetization distribution is based on solving inverse problem. Because the number of measurement points of magnetic flux density is different from the number of cells in which the magnetization vector is defined, singular decomposition is additionally required in solving inverse problem. On the other hand, when the target to minimize least square of difference between measurement flux and estimated flux is set, the unconstraint optimization problem can be formulated. Therefore, nonlinear programming can be applied to minimization problem of objective function. In this paper, the identification method based on quasi-Newton method or Newton method is proposed. The numerical performance of proposed method is demonstrated on practical permanent magnet which is employed on interior permanent magnet motor.

In sheet forming, the blank holder force (BHF) has a direct influence on product quality. For successful sheet forming, the BHF should be adjusted. Recently, the variable BHF (VBHF) that the BHF varies through stroke is recognized as one of the advanced sheet forming technologies. On the other hand, slide velocity (SV) that controls the die velocity is rarely discussed in the literature, and the SV should also be taken into account for the successful sheet forming. A high SV can achieve the high productivity, but wrinkling occurs. In this study, a design optimization approach using machine learning technique is adopted to determine them for achieving the high productivity. Numerical simulation in sheet forming is so intensive that a sequential approximate optimization using radial basis function network is adopted. Based on the numerical result, the experiment is carried out to examine the validity of proposed approach.

The present study optimizes lay-up configurations and drop-off placements for tapered laminates to maximize the failure strength. A lookup table of stress components for calculation of Christensen’s criteria is prepared in advance for all possible combinations of 3-ply laminates which is the minimum structure of ply drop-off. Stress components at three corners of triangle resin pockets are first obtained from the finite element analysis (FEA) for calculation of failure criteria. Then, the optimization is conducted for the laminate which tapers from 16- to 8-ply by using a simple genetic algorithm method. Objective functions are evaluated by referring only stress indexes listed in the lookup table and the FEA are not performed during the optimization. Both thick and thin sections are limited to quasi-isotropic laminates with 45° increment angle, but a stacking sequence of plies is variable and optimized. Placement of ply drop-off are also design variables. It was revealed that 0° and 90° plies are dropped near the thick and thin sections for the optimum result. Experimental results validated that the optimum structure has higher failure strength than the reference one.

In plastic injection molding (PIM), it is important to determine the optimal process parameters such as melt temperature and packing pressure for minimizing the warpage and cycle time. This work proposes a method to determine the optimal process parameters. Variable packing pressure profile that the packing pressure varies during the packing phase is used. Numerical simulation in the PIM is so intensive that a sequential approximate optimization is used to identify the pareto-frontier. The experiment is carried out to confirm that the proposed approach is valid to the warpage and the cycle time minimization.

Multiple tuned mass dampers (TMDs) are known to be effective when the system is excited by a wideband random disturbance. We study an influence of placement of multiple TMDs in spatial and frequency domain. In this problem, we decide the spatial arrangement, frequency and damping characteristics of the TMDs. Furthermore, we should consider multiple mode excitation close to the natural frequency in spatial structures. We present a method which is guaranteed to find a near-optimal solution with pre-assigned accuracy. The key concept of the method is use of order statistics with random search, which is also known as distribution-free tolerance intervals. Heuristic techniques: model order reduction techniques work effectively in conjunction with the proposed method without lack of validity of the theory.

The aim of this study was to formulate the grasping comfort of screw driver considering the individual difference of participants and optimize its shape. The shape parameters used in this study were the grip length L, end diameter d_e, and middle diameter d_m. The subjective scores of grasping ease were measured with 12 subjects. The relationship between the grasping comfort and shape parameters was modeled by the hierarchical Bayesian modeling. Regression coefficients were expressed by a combination of a global term for the whole participants and an individual term of each participant so as to consider the individual difference. Optimal shape parameters were determined by the predicted regression equation of grasping comfort, and the posterior distribution of grasping comfort for the optimal shape was investigated.

We have considered structure of large generator by parametric optimization for weight saving and stiffness improvement until now. However, the effect was often small because drastic improvement is generally difficult by parametric optimization. Therefore, we're studying reinforcement structure of generator using topology optimization which is non-parametric optimization. This paper shows a result of finding structure of bearing bracket whose eigen frequency exceeds 35 Hz by topology optimization. In addition, we carried out modal tests on conventional bearing bracket and optimized bearing bracket, and clarified eigen frequency improvement effect. We prove efficacy of topology optimization in design of large generators.

In this study, three dimensional running simulation of unilateral transfemoral amputee with running specific prosthesis was modeled. Based on motion capture data, kinematic and dynamic calculations are carried out to establish inverse dynamic model of human and prosthesis. Finally, prosthesis is optimized by using response surface method of inverse dynamic model.

When it comes to arranging the marine logistics infrastructure considering the economic growth of the region, one of the methods is adding shipping services. In this case, it is necessary to consider improvement of profit of shipping company operating the service and regional benefit, and we present it as an application example of the multiobjective function optimization method.

クレーンアタッチメントの構造という変数が多大であり、かつ複数の目的関数を持つような複雑な問題に対して、最適化手法を用いる範囲を限定し、問題の条件に合った最適化手法を用いることで、効率的に最適化を行い最良となる形状へと徐々に近づけていく手法を提案する．この論文では本手法が有効であるかの検証を行っていく．

A new framework of robust design is presented based on the concept of order statistics. A new multiobjective optimization problem is formulated to obtain solutions with various levels of robustness. A numerical example of a 20-story shear frame is presented for minimizing the maximum interstory drift angle under constraint on the total amount of damping coefficient, where uncertainty is considered in the story stiffness of the frame.

The steel house construction method using thin walled structure for the pillars and beams has attracted attention, because development of safer house has been required in recent years. For the thin walled structure, lighter weight and higher buckling strength are required simultaneously, while multiple buckling modes has to be considered. In this study, we made an optimum design program aimed at improving the buckling strength of lip groove-shaped steel combining Differential Evolution (DE) and Finite-Strip Method (FSM), and applied it to the buckling problem with compressive load and bending moment, to examine its effectiveness.