摘要:

In nonlinear optimal control problems, open-loop solutions from a fixed initial condition are much easier to compute than closed-loop solutions which do not depend on initial conditions. Two methods of using neural networks to approximate the optimal feedback controller are discussed. The indirect method uses a neural network to interpolate the whole field of extremals obtained from open-loop calculation. The direct method directly trains a neural network such that a general nonlinear optimal control performance index is minimized. The novelty of the modified backpropagation training is the requirement of the jacobian matrix of the neural network function. Simulation studies show that the closed-loop solution can be made to be arbitrarily close to the optimal open-loop solution with initial conditions chosen from a nontrivial subset of the state space.

ISSN:0020-7721    

DOI:10.1080/00207729408949275

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

A unified study of adaptive control and neural network based control schemes for the trajectory tracking problem of robot manipulators is presented. Efficacy of parametrized adaptive algorithms in compensating the structured uncertainties in robot dynamics is verified through extensive simulation. The ability of neural networks to provide a robust adaptive framework in the presence of both structured and unstructured uncertainties is investigated. A case study is carried out in support of a parametrized adaptive scheme using neural networks. Simulation results clearly indicate that the neural network based adaptive controller achieves better tracking in the presence of parametric uncertainties as well as unmodelled effects compared to the simple direct adaptive scheme.

ISSN:0020-7721    

DOI:10.1080/00207729408949276

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

Traditional optimal linear design problems are formulated by linear programming with a single criterion (objective)and a single resource availability (right-hand side) level. This approach is to seek an optimal solution for a given design system. However, it fails to deal with optimal linear design problems with multi-criteria and multi-source availability levels. In this paper, we first sketch how to use the multi-criteria and multi-constraint levels (MC2) linear programming to formulate the optimal linear design problems and to find a set of potentially good designs (PGDs). Then we generate generalized good designs (GGDs) from given PGDs and show how to construct their related dual contingency plans under various decision situations, fn contrast to the known concept of PGDs, the GGDs go beyond the limitations of the design opportunities. In constructing the dual contingency plans for a given GGD, we adjust the unit contribution of selected design opportunities in the GGD to convert non-optimal solutions into optimal solutions. Based on theoretical results, an algorithm of effectively and systematically locating the optimal generalized designs and the corresponding optimal dual contingency plans under various decision situations is provided.

ISSN:0020-7721    

DOI:10.1080/00207729408949277

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

Some results for checking the Schúr stability of uncertain polynomials are presented. In particular, it is shown that if the zeros of all the vertex polynomials are distributed in a certain fashion, then we can ascertain the Schur stability of the whole family of uncertain polynomials.

ISSN:0020-7721    

DOI:10.1080/00207729408949278

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

The paper considers two problems. Firstly the nonlinear localization problem; on the basis of available observations on the diffusion process, recover the location of an unknown single source generating this process. In general, the solution of this problem is set-valued and disconnected. Secondly, the identifiability problem; what types of observations (finite-dimensional at every moment of time, as generally occurs in physical situations) are able to provide enough information to restore the location point? A new approach is given, based on the introduction of a space of test-functions: in order to determine the unknown location, one needs to analyse a relevant system of algebraic equations. The latter can be determined in advance (‘off-line’). Sufficient conditions of identifiability are derived and the duality relations between the above nonlinear localization problems and the associated adjoint linear control problems are established. The above problems are motivated by environmental monitoring.

ISSN:0020-7721    

DOI:10.1080/00207729408949279

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

Robust adaptive control with rejection to unmodelled dynamics, unknown deterministic disturbance and other bounded disturbances using a new discrete time operator are discussed. It is shown that the adaptive version of the internal model principle can be combined with the relative dead zone to handle more general robust adaptive control problems. With the new operator, it is shown that the design of robust adaptive controller can be simplified and the stabilizability problem of the estimated model can be overcome. Global convergence and stability results are also established for the system without the requirement of persistent excitation. The analysis is supported by simulation results.

ISSN:0020-7721    

DOI:10.1080/00207729408949280

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

Some control-theoretic methods for evaluating the efficiency of production units in dynamic situations are developed and empirically illustrated on the basis of adjustment cost theory.

ISSN:0020-7721    

DOI:10.1080/00207729408949281

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

Continuous time mathematical models naturally have errors to be tolerated. Since the design of a compensator is usually based on the nominal values of a mathematical model, it is very important to determine the largest specified stable characteristic polynomial set centred on a given nominal characteristic polynomial in order to assess the robustness of the compensator, This paper presents a procedure to obtain the largest stable L, polynomial set centred on a given nominal characteristic polynomial.

ISSN:0020-7721    

DOI:10.1080/00207729408949282

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor

摘要:

A dual problem is proposed for an optimal stochastic control problem whose dynamical system is described by a linear stochastic differential equation. Relationships between the extreme values of the original and dual problems are discussed and two duality theorems are obtained The dual problem furnishes lower bounds for the extreme value of the original problem.

ISSN:0020-7721    

DOI:10.1080/00207729408949283

出版商:Taylor & Francis Group    

年代:1994

数据来源: Taylor