摘要:

An hierarchical control system is a control system in which there are several controllers of which some override or supervise the action of others. The aim of this paper is to show how a mathematical theory of problem decomposition can provide the basis for studying such systems. Previous writers have shown how control problems can be decomposed, but they have not shown how to obtain an hierarchy in which some subproblems have an overriding constraint on others. How this can be achieved is explained in the paper. It is demonstrated that one must introduce simplifications when decomposing the overall problem if one wants to obtain an hierarchy. In order to show how a mathematical theory of problem decomposition can be used to help analyse and design hierarchical systems, the decomposition process has been analysed in terms of a sequence of basic manipulations. This approach has several advantages. It has provided a clear understanding of the similarities and differences between established methods of decomposition, and has aided the development of a method for obtaining a problem hierarchy. Based on this analytical approach a new method for designing hierarchical control systems has been proposed. Instead of first specifying the ‘ low-level/layer ’ problems and later synthesizing ‘ high-level/layer’ problems, the new method begins by focusing on the overall control problem. An advantage of this method is that one is encouraged to recognize any simplifications introduced when decomposing the overall problem into subproblems

ISSN:0020-7179    

DOI:10.1080/00207177908922740

出版商:Taylor & Francis Group    

年代:1979

数据来源: Taylor

摘要:

This paper illustrates three applications of the decomposition method proposed in the preceding paper (Wilson 1979).. The first application illustrates how a multilevel hierarchical control system can be designed using the parametric decomposition and separation manipulations. The latter two examples illustrate the design of a multilayer control system using parametric and structural decomposition respectively.

ISSN:0020-7179    

DOI:10.1080/00207177908922741

出版商:Taylor & Francis Group    

年代:1979

数据来源: Taylor

摘要:

A system is considered in which a mathematical model contains a finite number of parameters which may vary or are uncertain as to value. Parameter dependence is then viewed as an operator valued map from a finite dimensional parameter space. This in turn leads to a design problem wherein the inverse to a finitely restricted system map is constructed. The potential applications of the design methodology include systems undera prioriand/or optimal control.

ISSN:0020-7179    

DOI:10.1080/00207177908922742

出版商:Taylor & Francis Group    

年代:1979

数据来源: Taylor

摘要:

The multiple-loop self-oscillating adaptive (SOAL) system is presented as a natural, logical means of overcoming a serious limitation of the single-loop self-oscillating system (SOAS). Both structures have the property of zero sensitivity to plant high-frequency gain uncertainty p = Kmax/Kmin, the factor which is generally responsible for large ‘ cost of feedback ’. It is however necessary to design these systems such that the response is essentially quasi-linear to the desired class of command and disturbance signals. In the SOAS,preappears as a significant factor in the quasi-linear requirements which may, depending on the numbers involved, completely vitiate its banishment as an uncertainty factor. The development of aquantitativedesign theory for the SOAS pinpoints the two-loop SOAL extension needed to overcome this SOAS limitation, and the development of a similar SOAL quantitative design theory. In the latter,pdisappears from both the adaptive and quasi-linear conditions, but reappears as a factor in the rate of adaptation of the system. It may be banished from here too, by means of a three-loop self-oscillating system (SOANL), for which the SOAL design theory is applicable with minor extensions.

ISSN:0020-7179    

DOI:10.1080/00207177908922743

出版商:Taylor & Francis Group    

年代:1979

数据来源: Taylor

摘要:

It has been known for some time that modifications may be necessary to existing conditions at junctions on partially singular trajectories. Such modifications would be governed by the role played by certain time derivatives which exhibit discontinuities at the junction. This paper investigates the simplest case in which such discontinuities can appear and demonstrates that either the modifications are not necessary or that the usual analysis is not sufficient to establish a result. An example is given to illustrate one of the existing conditions necessary at a junction.

ISSN:0020-7179    

DOI:10.1080/00207177908922744

出版商:Taylor & Francis Group    

年代:1979

数据来源: Taylor

摘要:

Necessary and sufficient conditions are determined for the block decoupling of an J-input, m-output, linear, time-invariant and minimal system (where each block contains more than one output) using state feedback and a constant non.singular input transformation. A technique for solving the block decoupling problem is also proposed.

ISSN:0020-7179    

DOI:10.1080/00207177908922745

出版商:Taylor & Francis Group    

年代:1979

数据来源: Taylor

摘要:

A series of controlled jump process models is presented for stochastic scheduling problems. In each case, the optimality principle of dynamic programming is used to reduce the problem to a deterministic optimal control problem. Applying the maximum principle to this deterministic problem yields necessary optimality conditions. These enable explicit optimal scheduling strategies to be found for a number of problems.

ISSN:0020-7179    

DOI:10.1080/00207177908922746

出版商:Taylor & Francis Group    

年代:1979

数据来源: Taylor

摘要:

It is shown that the [n(n+ l)/2] conditions for stability in the left-half plane as well as inside the unit circle as given by Routh and Jury-Gutman, can be reduced reapectively to {[n[n− 1)/2] + l} and {[n[n− 1)/2] + 2} conditions. Furthermore, a counter example of sixth degree polynomial to Fuller's conjecture of n-conditions is obtained. Finally, methods for obtaining polynomials for root-pair-sums, root-pair-product and polynomials having as roots the negative of squares of root-pair-differences (needed for aperiodicity conditions are obtained.

ISSN:0020-7179    

DOI:10.1080/00207177908922747

出版商:Taylor & Francis Group    

年代:1979

数据来源: Taylor

摘要:

The design of proportional-plus-integral compensators for linear multivariable systems is considered, using an output feedback control law. An algorithm is developed based on results obtained elsewhere (Munro and Novin-Hirbod 1979). It is shown that cyclicity of the augmented system matrix is not needed with this approach. It is also demonstrated that the integral part of the compensator should be of full-rank for an arbitrary set of non-zero eigenvalues to be assignable.

ISSN:0020-7179    

DOI:10.1080/00207177908922748

出版商:Taylor & Francis Group    

年代:1979

数据来源: Taylor

摘要:

The problem of output feedback controller design with low trajectory sensitivity to small system parameter variations is considered for both deterministic and stochastic systems. The performance index is of the standard linear regulator type modified to include a quadratic sensitivity term. A concise derivation of the necessary conditions for optimality is presented for deterministic systems using the method of Lagrange multipliers and existence of an optimal solution is proven. The low sensitivity controller is realized by a gain matrix obtained from the solution of a set of non-linear matrix equations. Also, use is made of recent observer-estimator results to present a unified approach to low-sensitivity controller design for stochastic systems.

ISSN:0020-7179    

DOI:10.1080/00207177908922749

出版商:Taylor & Francis Group    

年代:1979

数据来源: Taylor