In our previous work on supervisory control of discrete event systems, we proposed a novel supervisor which assigns a control pattern based on partial observations of both events and states, and showed necessary and sufficient conditions for its existence. This paper extends the result to decentralized supervisory control of large-scale discrete event systems consisting of several subsystems which operate concurrently. We address two types of decentralized control problems. One requires that the behavior of the closed-loop system equals a given legal language. The other requires that the behavior of the closed-loop system lies in a given admissible range. For each problem, we derive a necessary and sufficient condition for the existence of a decentralized supervisor.
This paper presents necessary and sufficient conditions for a linear descriptor system to be quadratically stabilizable via linear feedback control. Uncertainties in the system are structured and norm-bounded. It is assumed that the system has no impulsive mode and the coefficient matrix for the derivative of the descriptor variable has no uncertainty. Quadratic stabilizability is defined by introducing a positive semidefinite quadratic function which is, however, positive definite for dynamic behaviors of the descriptor variable. The stabilizability conditions are described in terms of the existence of certain solutions of a generalized algebraic Riccati equation.
The realization of a shape memory alloy (SMA) actuator has been demanded in the field of microactuator for a long time. But it is difficult to control a SMA actuator precisely because of its non-linear hysteresis properties. To compensate this hysteresis property, we propose a control system in this paper. At first, we design a model of hysteresis properties and its inverse model. Then, we propose a controller to compensate the hysteretic properties by using this inverse model. Finally, we apply this method to control SMA resistance and the usefulness of this controller was demonstrated experimentally.
In the external beam radiotherapy treatment of cancer, it is important to plan the treatment so that the irradiated energy the tumor is sufficiently high while the dose level at the normal tissue is made as small as possible. In planning a treatment with multiple beams, it is necessary to optimize the position and the intensity of the sources of radiation. In this study, we proposed a treatment planning method by using Mixed Integer Programming (MIP), which can control the number of beams, comsidering the operational feature of the treatment. The effectiveness of the proposed method is demonstrated by an illustrative case study.
The important feature of the projection pursuit (PP) is that it is one of the multivariate methods able to bypass the “curse of dimensionality”. The aim of PP is to find an interesting or characteristic structure by working in low-dimensional linear projections. PP for regression was originally proposed by Friedman and Stuetzle. In this paper, a neuro-fuzzy approach to the projection pursuit regression (PPR) is proposed for nonparametric regression and nonparametric classification. Our proposed method is based on the membership function and the eigenvector of the covariance matrix to avoid being trapped by a local minimum of the projection indices. The dimensionality of predictors is reduced to one or two in order for good use of human ability of instantaneous pattern discovery. The radial basis function neural network is applied to the function approximation in a projected low-dimensional space. The projection direction is also changed by the adaptive learning (steepest descent) method.