2014-07-09 - Two papers accepted at ETFA 2014 conference

The papers:

"Enhancing Banker's Algorithm for Avoiding Deadlocks in Systems with Non-sequential Processes" that has been presented with Dr. Michal Žarnay from the University of Žilina.

Abstract:
Deaddlock states are undesirable situations in any system. Their avoiding presents a major issue in control of resource allocation systems (RAS), a theoretical paradigm used in construction of several application systems, including flexible manufacturing systems. In coordinating RAS with non-sequential processes, i.e. processes containing concurrent execution in their development, it has been proved that deadlock avoidance policies developed for sequential systems are applicable. This requires, however, transformation of the coordinating RAS structure using reachability graph of single processes to a sequential RAS, what brings the exponential growth of the modelled system size.
In this paper, we present a deadlock avoidance method that allows using the original system, coordinating RAS with non-sequential processes, with additional data structures calculated before the system start. It is a banker's like approach -- an enhanced version of the banker's algorithm with our proposed data structures manipulation.


"The deadlock problem in the control of Flexible Manufacturing Systems: An overview of the Petri net approach" was presented with Drs. J. P. López-Grao and J. M. Colom from the Universidad de Zaragoza.

Abstract:
The disparity of Petri net models in the literature for the study of resource allocation problems in Flexible Manufacturing Systems (FMSs) is overwhelming. Paradoxically, those models often strongly overlap in terms of their modeling capability. Such clustering seems to be impelled by a mature knowledge of results which capsize liveness at the topological level of the FMS for families of models which comply with some properties of good behaviour. Not rarely, such properties are embraced at the expense of severe or inconsistent syntax restrictions paying slight regard to actual modelling requirements in the application domain of FMSs. In this paper, we promote a distillation of the current amalgam of subclasses and tangent results, proposing taxonomic categories that allow capturing the different capabilities of each kind of model and establishing a reference framework for past and future advances.
Escrito el 2014-07-09
Categorías: research