In the fuel supply facilities where fuels are unloaded from tankers to tanks and supplied through pipelines, reducing the number of valve operations is important to improve the efficiency of facility utilization. This is a combinatorial optimization problem that asks to minimize the number of changing the assignment patterns of tank groups to berths. We propose here an effective heuristic algorithm which exploits the special structure of the problem. Computational results indicate that this can be used as a scheduler in real-world supply facilities.
The time tabling problem is an important combinatorial optimization problem that arises in schools and so on. In this paper, we treat the problem of time tabling by incorporating the requests. This kind of problem occurs when the problem is commercially-based so that the applicants' requests must be highly satisfied, or many of the applicants are so busy that they can be assigned only to the time slots which they request. For such problems, we propose an approach using the genetic algorithm (GA). We formulate this problem as the permutation problem of the order of assignment of the requests, where the permutation is defined as the chromosome of the GA code. We circumvent the hard constraints by using the allocation module, and try to pursue the local optimality by using the minor tuning module.
This paper presents a predicting method for standard deviations of chemical composition fluctuations of raw meal at a kiln inlet using chemical analysis results of boring samples for raw materials in a quarry. A stochastic process representation for their fluctuations at the inlet of the plant is presented to specify their time domain fluctuations. The authors have also presented dedicated transfer functions which represent homogenizing effect of each constituent process and the raw material mixing control system. Using these representations, the Bode-diagrams to obtain decreasing performances of the fluctuations are finally presented. The theoretical calculation result of the standard deviation is compared with the result measured in an actual plant. As a result of a succession of these analyses, moreover, the authors could also present the structure of the mixing control system as a multivariable stochastic control system.
It is known in the physiology that the apoptosis is an active form of cell death in most multicellular organisms and one of the two mechanisms by which cell death occurs (the other being the pathological process of necrosis). Apoptosis is the mechanism responsible for the physiological deletion of cells and appears to be intrinsically programmed. We propose a procedure called M-apoptosis for the structure clarification of Neurofuzzy GMDH model whose partial descriptions are represented by the Radial Basis Functions network. The proposed method prunes the larger network to identify, still more to clarify the network structure by minimizing the Minkowski norm of the gradient with respect to input variables of the partial descriptions. The method was validated through graphical representations of the identified structures in the numerical example of function approximation and classification of Iris data.
The common Lyapunov function problem arises in association with stability analysis of diverse fields of systems. The problem is numerically solvable. However, it is not an easy task to fully characterize such a class of systems which have a common quadratic Lyapunov function. It is thus far known that a set of linear stable systems has a common quadratic Lyapunov function if their system matrices are in a commuting family. But this condition is considerably restrictive. Our objective here is to find another class of systems which has a common quadratic Lyapunov function. It turns out that if all systems have stable system matrices which have an upper (lower) triangular structure, then they have a common quadratic Lyapunov function. Note that such system matrices are not necessarily in a commuting family. The obtained results would give an insight into further exploration of the common Lyapunov function problem. Numerical examples are worked out for illustration.
In the preceding paper, experimental results on the control of a magnetic levitation system using the ILQ and the H∞. control theories were reported. In the ILQ case, a spillover was caused by unmodeled dynamics, where the dimention of the controller was lower than that in the H∞ case. In this paper, we shape the closed loop frequency response of the ILQ control system to suppress the spillover by introducing additional design freedom of the observer. As a result almost the same closed loop performance as the H∞ controller can be achieved by this lower dimensional controller.