Steepest-descent optimal control techniques have been used extensively for dynamic systems in one independent variable and with a full set of initial conditions. This paper presents an extension of the steepest-descen...
详细信息
Steepest-descent optimal control techniques have been used extensively for dynamic systems in one independent variable and with a full set of initial conditions. This paper presents an extension of the steepest-descent technique to mechanical design problems that are described by boundary-value problems with one or more independent variables. The method is illustrated by solving finite-dimensional problems, problems with distribution of design over one space dimension, and problems with distribution of design over two space dimensions.
This short paper studies a particular class of optimization problems dealing with the selection, at each instant of time, of one out of many actuators in order to obtain a determined result. A cost is associated with ...
详细信息
This short paper studies a particular class of optimization problems dealing with the selection, at each instant of time, of one out of many actuators in order to obtain a determined result. A cost is associated with each actuator. The cost function is the integral of a weighted combination of the achieved accuracy on the state of the system and the control energy. The control energy term depends upon both the selected actuator and the magnitude of the applied control. The problem is to design an optimal actuator selection strategy. The analysis is limited to the class of linear deterministic systems with measurable states. A discrete approach is considered. The analytic solution to this optimization problem is given first. When the number of actuators and the number of stages in the time interval become large the optimal analytic solution requires a considerable combinatorial work; a suboptimal algorithm is then proposed to alleviate this defect.
The methods of system identification due to Shinbrot, Perdreauville and Goodson, Fairman and Shen, and Diamessis belong to a class of techniques based on the idea that a linear operation on system equations leads to a...
详细信息
The methods of system identification due to Shinbrot, Perdreauville and Goodson, Fairman and Shen, and Diamessis belong to a class of techniques based on the idea that a linear operation on system equations leads to a set of simultaneous equations solvable for the unknown parameters. These methods have so far been able to treat problems of linear time invariant, time-varying, nonlinear, and distributed parameter systems without lags. This short paper extends the capability of the class of methods to cover time-lag systems. The extension is illustrated with reference to one of the methods that employs Poisson moment functionals. Illustrative examples include linear time invariant, time-varying, and non-linear systems each with a lag.
Necessary conditions are derived for the optimum control of a class of smooth second-order distributedsystems subject to general boundary conditions of the Neumann type.
Necessary conditions are derived for the optimum control of a class of smooth second-order distributedsystems subject to general boundary conditions of the Neumann type.
In this note it is shown that if a distributedparameter system of parabolic type is controllable by distributed controls then it is also controllable by a scanning control, i.e., a single point control that can be am...
详细信息
In this paper an equivalent one-dimensional system for taper leaf springs (similar to those used in vehicle suspension systems) is developed so that results obtained from it are in good agreement with those obtained f...
详细信息
A method is presented for identifying linear distributed parameter systems. Emphasis is placed on identification as a function of spatial coordinates by considering time-transformed, noise-free systems. Measurements o...
详细信息
A method is presented for identifying linear distributed parameter systems. Emphasis is placed on identification as a function of spatial coordinates by considering time-transformed, noise-free systems. Measurements of system response are combined with the Green’s function method of analysis to obtain integral equations that can be solved for unknown spatial operators or coefficients. A discrete form of the theory is developed, utilizing Chebyshev polynomials. This allows prior estimates to be used to determine the number and location of spatial measurements. Where estimates are of sufficient order, the modeling process is exact.
The problem of matching a linear time-delay system to a given time-delay model is considered and solved in state-space form. The system under control is assumed in its phase-variable form. A two-dimensional example is...
详细信息
The problem of matching a linear time-delay system to a given time-delay model is considered and solved in state-space form. The system under control is assumed in its phase-variable form. A two-dimensional example is worked out to illustrate the solution method.
暂无评论