Maria Elena Meda-Campaña

Incremental synthesis of Petri net models for identification of discrete event systems

This paper addresses the problem of online identification of discrete event systems (DES). A passive method for the progressive building of Petri net (PN) models from DES outputs evolution is presented. After introducing several concepts related with dynamical properties of DES, a learning algorithm that computes ordinary PN models according to the measurement of cyclic output streams is proposed. A procedure based on this algorithm can be on-line executed tracking the DES behavior from its output signals; the successive computed models tend progressively to represent the actual observed behavior.

Required event sequences for identification of Discrete Event Systems

In previous works was addressed the on-line identification problem. This problem consists in compute an Interpreted Petri Net (IPN) model in proportion as new input and/or output sequences of the system are observed. Now in this paper are presented the required transition sequences needed to identify an IPN model that describes the complete behavior of a DES. These transition sequences are important because if from the knowledge of the output signals of the system we can compute these kind of transition sequences, then it will be guaranteed that the complete behavior of the system is captured in the computed model, even not all transition sequences of the system have been detected.

Asymptotic identification of discrete event systems

This work deals with the analysis of the online identifiability of the discrete event systems (DES) using interpreted Petri nets. A theoretical framework characterizing identifiable systems when only the input and output systems signals are available is first addressed. Based on this framework, an algorithm that progressively builds an IPN model of the system is then presented. As a possible application of the problem herein addressed, the online identification problem is adapted to address the problem of DES model validation.

Fault detection and location in DES using Petri nets

This paper is concerned with fault detection and location of discrete event systems (DES) modeled using interpreted Petri nets (IPN). The approach held deals with IPN models in which the marking is partially known. A diagnosis scheme that tracks the DES inputs and outputs is proposed; it allows detecting and locating an unexpected behavior within the DES. The diagnoser handles a monitoring model, expressed also as an IPN, which is computed from the DES model through an induced conservative marking law.

Synthesis of timed Petri net models for on-line identification of Discrete Event Systems

This paper addresses the problem of on-line identification of Discrete Event Systems (DES). A passive method for the progressive building of Petri net (PN) models from DES outputs evolution is presented. After introducing several concepts related with dynamical properties of DES, a learning algorithm that computes ordinary PN models according to the measurement of cyclic output streams is proposed. A procedure based on this algorithm can be on-line executed tracking the DES behavior from its output signals, whose durations are stored. The successive computed models tend progressively to represent the actual observed behavior.

Use of evolved artificial regulatory networks to simulate 3D cell differentiation

Cell differentiation has a crucial role in both artificial and natural developments. This paper presents results from simulations in which a genetic algorithm (GA) was used to evolve artificial regulatory networks (ARNs) to produce predefined 3D cellular structures through the selective activation and inhibition of genes. The ARNs used in this work are extensions of a model previously used to create 2D geometrical patterns. The GA worked by evolving the gene regulatory networks that were used to control cell reproduction, which took place in a testbed based on cellular automata (CA). After the final chromosomes were produced, a single cell in the middle of the CA lattice was allowed to replicate controlled by the ARN found by the GA, until the desired cellular structures were formed. Two simple cubic layered structures were first developed to test multiple gene synchronization. The model was then applied to the problem of generating a 3D French flag pattern using morphogenetic gradients to provide cells with positional information that constrained cellular replication.

A passive method for online identification of discrete event systems

This paper addresses the problem of online identification of discrete event systems (DES). The approach held is a passive one: an interpreted Petri net (IPN) model is built from the measuring of the DES output evolution. First, several behavioral properties of DES are introduced, then an algorithm that computes IPN models from cyclic output streams is presented. In an online operation the currently obtained model represents the observed behavior of the system and successively computed models lead asymptotically to the actual one.

Use of a Feedforward Neural Network for Predicting the Development Duration of Software Projects

Context: In the software engineering field, only 20 percent of software projects finish on time relative to their original plan. A software project can be classified as a new development, an enhanced development or a re-development. Goal: To propose a feed forward neural network (FFNN) for predicting the duration of new software development projects. Hypothesis: The accuracy of duration prediction for an FFNN is statistically better than the accuracy obtained from a statistical regression (SR) when an adjusted function points (AFPs) value, obtained from new software development projects, is used as the independent variable. Method: A sample obtained from the International Software Benchmarking Standards Group (ISBSG) Release 11 corresponding to new development projects was used. The accuracy of the FFNN was compared against that of an SR model. The criteria for evaluating the accuracy of these two models were the Mean Magnitude of Relative Error (MMRE) and an ANOVA statistical test. Results: Prediction accuracy of an FFNN was statistically better than that of an SR model at the 90% confidence level. Conclusion: An FFNN could be applied for predicting the duration of new software development projects when AFPs were used as independent variable.

Software Development Productivity Prediction of Small Programs Using Fuzzy Logic

In this paper, a fuzzy logic model was created from a data set of 140 small programs developed with practices based on Personal Software Process (PSP) and then this fuzzy model was applied for predicting the productivity of a new data set consisted of 60 small programs; all programs were developed with object oriented programming languages by 35 and 15 graduated programmers respectively. Accuracy result of this fuzzy logic model was compared with that of a statistical regression model. Results suggest that a fuzzy logic model could be used for estimating and predicting productivity of the software development.

Dynamical local properties for estimation and control of discrete event systems modeled by interpreted Petri nets

This paper studies the properties of observability, controllability and identification in discrete event systems (DESs) modeled with interpreted Petri nets (IPNs). It is important to note that these properties are defined using the states of the IPN and use structural information to characterize IPNs that exhibit these properties. This is an improvement on previous works because these definitions allow one to study more powerful control schemas.