Performance evaluation and perturbation analysis of discrete event dynamic systems (Q3763989)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 4020953
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | Performance evaluation and perturbation analysis of discrete event dynamic systems |
scientific article; zbMATH DE number 4020953 |
Statements
Performance evaluation and perturbation analysis of discrete event dynamic systems (English)
0 references
1987
0 references
Discrete event dynamic systems
0 references
perturbation analysis
0 references
0.8672299981117249
0 references
Discrete event dynamic systems (DEDS) are systems in which the state changes only at certain discrete instants of time, instead of continuously, depending on complex interactions of the timing of various discrete events occurring inside the system itself. Typical examples are man-made systems such as flexible manufacturing systems, communication networks, air traffic control systems etc. These systems are assuming an oustanding interest and are drawing an increasing attention. The paper seems to be successful in motivating the importance of the topic in question and in calling for research efforts to develop a theory and methodologies, that can parallel those established for continuous-time systems. In fact, the theory of DEDS is not as developed and understood as the theory of continuous dynamical systems, and the numerical tool-box for solving estimation and control problems is far from being satisfactory. The paper surveys, at an intuitive legel, various approaches to the analysis of these class of systems with particular emphasis on perturbation analysis. This emphasis is justified by the author with his bias against computationally unfeasible tools. Actually it is usual, when there are no better methodological tools, resort is made on simulation and perturbation techniques. However, for the case of DEDS to carry on perturbation analysis becomes a nontrivial task essentially because of the discontinuities that occur in these systems. Here, in my opinion, the points that the author tries to make, could be really appreciated by the interested reader, only if a precise specification of the underlying function (sample) spaces were given, along with their topological and probability measure structure. My impression is that this field awaits an approach that is at a time satisfactory at theoretical level, realistic at modelling level, and viable at numeric level or at least able to assess separately the complexity of each method and problem. In fact, intractability may well be an intrinsic feature of the problem, and we should ascertain in which cases that is so. To achieve advances in this direction some fundamental points should be addressed. I strongly agree with the author's comments that a dynamical analysis is unescapable. But then the first modelling problem we have to address regards the probability spaces to be adopted. In this respect there may well be cases, in which continuous variables are not adequate, in the same way as today we would not try to approximate an integer (or mixed integer) programming problem with a continuous programming problem. In this spirit I appreciate the author's suggestion that a separate paper addressing pure theory should be published, but I would take this suggestion one step further. Because of the interest of the topic in question and because so many interesting approaches are only touched upon, while other more recent contribution are emerging in these days, why not devoting a whole issue to the topic, which may give space to all existing approaches and may attempt a more complete assessment of the state of art. I am sure such an initiative would have a large and interested audience.
0 references