Grammatical Evolution (GE) is a kind of evolutionary computation which aims to design the function or program satisfying the design objective. The search algorithm of GE is same as Genetic Algorithm (GA). For improving the search performance, Grammatical Differential Evolution (GDE) is presented. GDE adopts the Differential Evolution (DE) instead of GA. This paper focuses on the improved algorithm of GDE, which is named as Real-value Grammatical Defferential Evolution (RGDE). GDE and RGDE are compared in the symbolic regression problem. The results show that RDGE has better convergence property than GDE.

This study focuses on the simulation of a vehicle leaving from the vehicle platoon of four vehicles. The velocity control model is defined by means of the improved Gazis model, which is one of the vehicle following models. The experiment of vehicle platoon simulation of four LEGO Mindstorms is performed. The results are compared with numerical simulation in order to confirm the validity of the model.

In this paper, we present global shape identification in incompressible viscous flow based on the fictitious domain FEM and the Kriging response surface approach. In the fictitious domain FEM, foreground and background meshes are employed. If this method is employed for the moving body problems, remeshing is not necessary. In this study, we apply the fictitious domain FEM to identify an airfoil shape in an incompressible viscous flow field based on the Kriging response surface approach.

In this study, we present numerical examples for estimation of shallow water flow based on the Kalman filter FEM. The shallow water equation is employed as the governing equation, and the Galerkin method using the triangular element and the selective lumping method are applied to discretize the governing equation in space and time, respectively. We investigated difference of numerical results by changing position of the observation points. It was found that high accurate flow field estimation result is obtained in case that the observation points are set to the upstream side in the open channel in comparison with the case that the observation points are set to the downstream side in the open channel. Therefore, inflow boundary condition is introduced to estimation analysis of shallow water flow based on Kalman filter FEM, and examinations of inflow boundary condition on the estimation accuracy are carried out.

In this study, we introduced the smoothing methods into the three-dimensional shape identification analysis based on thermal testing data and compared difference between computational results employing each techniques. Then two models, including smooth shape of an ellipse and other one with a dent, are given as target shape, and time history of the temperature on each target shape is obtained to apply to shape identification analysis. As a result, it is found that appropriate smoothing technique on the shape identification analysis changes according to conditions in target shape. Gaussian filter is effectual in a case where the shape of cavity is identified as including smoother surface from the aspect of performance function. On the other hand, in the case that the cavity surface of the target shape has a dimple, traction method has efficacy for bringing the computational model close to target shape.

The author received a question "What is solution growth?" from the editorial committee of Transactions of the Japan Society of Mechanical Engineers. Then, as the answer, this presentation introduces a process for fabricating light-emitting diodes, surveys the other fields such as applied physics and discusses the necessity of mechanical engineering.

The stress increment directing inwards of the current subloading surface does not always lead to an elastic unloading process without a plastic strain increment. This fact must be taken account for the loading criterion in the elastic predictor step of the return-mapping projection. The exact formulation for the loading criterion in the elastic predictor step will be given for the initial and the extended subloading surface models.

Data produced by finite element analysis (FEA) programs is increasing year by year, and data transfer and storage are significant issues for efficiency and productivity. Therefore, we propose a new method of data reduction for parallel FEA employing domain decomposition method (DDM). Based on the DDM and a data compression technique designed specifically for floating-point numbers, called jointed hierarchical precision compression number-data format (JHPCN-DF) which is available for lossless and lossy compression with user-defined errors, we develop adaptive JHPCN-DF for efficient visualization and analysis. In this study, we implement adaptive JHPCN-DF within a FEA program and evaluate not only the compression effectiveness but also the accuracy of compression data in real FEA. The results indicate that our method provides significant improvement of compression performance and achieve high accuracy.

The shape optimization has an effect of leveling stress distribution in the structures of many types, and the reduction of the peak stress amplitude has an influence on the improvement of its durability. The leveling effect of the shape optimization is investigated, on the stress distribution of the Michell truss under the dynamic loading with sweep excitation. The number of troughs and peaks can be divided into 2(two) groups, i.e. troughs and peaks with elastic strain only, and troughs and peaks with plastic strain accumulation, which leads us to apply the Palmgren-Miner rule directly. In this paper, the effect of leveling stress distribution in the Michall truss is reported, regarding the improvement of durability evaluation.