Optimization by Vector Space Methods
Optimization by Vector Space Methods
Hierarchical Strategies for the On-Line Control of Urban Road Traffic Signals
5th Conference on Optimization Techniques, Part 2
The modelling and control of water quality in a river system
Automatica (Journal of IFAC)
A hierarchical strategy for the temperature control of a hot strip roughing process
Automatica (Journal of IFAC)
Paper: Optimal control of constrained problems by the costate coordination structure
Automatica (Journal of IFAC)
Brief paper: Closed loop hierarchical control for river pollution
Automatica (Journal of IFAC)
Brief paper: The optimization of non-linear systems using a new two level method
Automatica (Journal of IFAC)
Hi-index | 22.15 |
This paper gives a brief survey of possible methods which can be used for the practical control of large interconnected dynamical systems. The development is in two parts, i.e. optimal methods and suboptimal methods. In the first part, a brief outline is given of infeasible methods like Goal co-ordination and the Takara-Sage algorithm. In the general study of infeasible methods, Tamura's three-level method, Tamura's time-delay method and Pearson's pseudo-model co-ordination method are also included. It is seen that both the algorithms of Tamura as well as the Takahara-Sage method are particularly suited to systems with slow dynamics whereas Pearson's pseudo-model co-ordination method could be used advantageously for systems with fast dynamics. A practical example is then given of optimal traffic control using an infeasible method, in this case the time-delay method of Tamura. The main conclusion to emerge from this part is that optimal methods will require multiple processors for on-line dynamic optimization although for systems with slow dynamics like the traffic example, fairly large problems could nevertheless be tackled using only one processor. There are certain classes of systems for which it may be possible to obtain virtually optimal control using only a single processor even when the number of subsystems is very large. One such class of problems is of serially connected dynamical systems. In the second part of this paper a suboptimal approach is described for the control of serial systems and the method is demonstrated using river pollution as an example. Finally, a new method is developed which enables a significant improvement to be made for serial systems with conflicts between the subsystems and an example illustrates this approach.