Ph.D. received on: 16/4/1998

E-mail: martinel@disp.uniroma2.it

Tutor: Prof. Salvatore Nicosia, University of Roma Tor Vergata
 
 

 ___________________________________________________________________________________________________________

Modeling and Control of Discrete Event Dynamic Systems   ___________________________________________________________________________________________________________

Advisor:

Prof. Salvatore Nicosia, University of Roma Tor Vergata
 

Summary:

The thesis deals with the problem of dynamic control of Discrete Event Systems.
Two main problems have been investigated: the estimate of a performance function defined on the system (Perturbation Analysis) and the dynamic resource allocation to optimize such a performance function.
A new algorithm is given in the thesis, which has been developed with the collaboration of Dr. Paolo Valigi (University of Roma Tor Vergata, Italy), after a comprehensive analysis of previous results, which benefitted also from the collaboration with prof. Christos Cassandras (Boston University, USA). Such an algorithm can be used to estimate a performance function defined on the system as a function of some parameters to be modified in order to optimize the performance of the system.
There are many advantages in the proposed algorithm with respect to other previous techniques: it can be applied to any discrete event system, with any perturbation on the parameters, providing an exact reconstruction of the state based on the information of the past of the system.
The dynamic choice of the parameter vector to optimize the performance function is the way the system is controlled. Also in this framework a new algorithm has been introduced. It is characterized by many interesting features of other control schemes but it can be applied to non stationary systems. The control of non stationary systems is both a challenging problem and very useful from a practical point of view.
A faster version of the algorithm is also offered in the thesis. The thesis contains analytical and simulative results about the algorithms and also comprises a comprehensive overview on previous results and techniques which are compared with the proposed algorithms.
 

  _______________________________________