[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

In this study, we propose an outer approximation method for a global optimization problem constrained by convex and concave functions. It is known that such a problem can be transformed into a d.c. programming problem. The proposed algorithm utilizes a sequence of convex polyhedral sets approximating the reverse convex set.

View full abstract

This paper proposes new optimal location problems considering preference of customers for competitive facilities, e.g. shops. In most problems, customers only use the facility with the highest valuation. However, customers sometimes use other facilities not much inferior. Problems considering such situations are formulated and an efficient solving algorithm based upon tabu search is proposed.

View full abstract

[in Japanese]

View full abstract

DEA is well known as the method to measure efficiency of DMU (Decision Making Unit) with multiple inputs and multiple outputs. Using concomitant variables, this paper shows that the basic DEA model (CCR) can be spread out into two DEA models: IRS model which suggests the input-based improvement proposal, and DRS model which suggests the output-based improvement proposal. And this paper proposes the framework which selects improvement direction mathematically form the relationship diagram between two DEA models.

View full abstract

The author proposed "Recurrent Evaluation Model" by which sustainable society can be designed. In this paper, the algorithms of directions, performances and evaluations by evaluators are clarified through the variational network analysis of phase plane characteristics used in the above model.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

Previous study showed that thermal sensation vote (TSV) affects thermal comfort vote (TCV). This paper discusses the comfortable preference structure based on Bayesian Network (BN). The proposed BN model has three layers; (1) change of TSV, (2) TSV, and (3) TCV. Numerical analysis based on 1,584 samples shows the causal relation among three layers.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

In this paper, we consider stability analysis and design for switched systems consisting of linear discrete-time descriptor subsystems. When all descriptor subsystems are stable, we show that if the descriptor matrix and all the subsystem matrices are commutative pairwise, then the switched system is stable under arbitrary switching. We also extend the result to the case where all subsystems have different descriptor matrices. Under the same commutation condition, we show that in the case where all the descriptor subsystems are not stable, if there is a stable combination of the unstable descriptor subsystems, then we can establish a class of switching laws which stabilize the switched descriptor system.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

In this paper, we are concerned with a problem of optimal selection of the gain matrix of the observation in linear stochastic systems. By introducing an information theoretic constraint based on a generalized Water Filling Theorem, we obtain a set of equations to obtain the optimal gain matrix which produces the best performance of the optimal LQG regulator.

View full abstract

In this paper, we are concerned with a problem of optimal selection of the gain matrix of a linear observation for the stationary Kalman filter. By introducing an information theoretic constraint based on a generalized Water Filling Theorem, we obtain a gain matrix which minimizes the stationary error variance.

View full abstract

We propose a new path control method with path error search and its application in this paper. We search the shortest distance point on a target path from a present position and the shortest distance is a essential path error. The shortest distance line is considered as a standard axis and coordinate transformations for a mechanical system are performed. An effective path control with little path error is realized by the proposed method.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

This paper proposes a new multi-objective design method of gain-scheduled controllers based on descriptor representation of LPV systems. We apply new LMIs for descriptor systems that have been shown recently to synthesis of gain-scheduled controllers under multiple specifications on control performance.

View full abstract

The problem of H-infinity norm minimization is represented as a feasibility problem of a bilinear matrix inequality, and the numerical optimization methods have been studied. The author has proposed a descent method. In this method, the bilinear matrix inequality is approximated by a linear matrix inequality, and the descent direction is calculated using the solution. Next, a line search method is applied to find a better controller. Numerical experiments show that the norm decreases quickly but after that it moves abruptly in some cases. In this paper, these phenomena are examined in detail, and a new line search method is proposed to improve the decrease.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

A new closed loop subspace identification method has been developed by using measurable disturbances. The present method is based on the orthogonal decomposition of control inputs and outputs onto the space of disturbance. We extend the results presented at SCI'06, and show some numerical simulations for SISO and MIMO systems.

View full abstract

[in Japanese]

View full abstract