The volume 2007

Table of Contents

Cornelia NOVAC-UDUDEC, Rodica BIRLA: A Student Model in a multi-agent intelligent system
Abstract  Full Text

Nicolae JASCANU, Veronica JASCANU, Florin NICOLAU: A new approach to e-commerce multi-agent systems
Abstract  Full Text

Adrian SERBENCU, Viorel MINZU, Adriana SERBENCU: An Ant Colony System based metaheuristic for solving single machine scheduling problem
Abstract  Full Text

Cornelia TUDORIE, Laurentiu FRANGU, Diana STEFANESCU: Automatic modeling of the linguistic values for database fuzzy querying
Abstract  Full Text

Adina COCU: Knowledge-based diagnostic system with probabilistic approach
Abstract  Full Text

Dan MUNTEANU: A quick survey of text categorization algorithms
Abstract  Full Text

Dan MUNTEANU: Vector space model for document representation in information retrieval
Abstract  Full Text

Ana-Elena LUNGU, Daniela COLTUC, Laurentiu FRANGU: Computing negentropy based signatures for texture recognition
Abstract  Full Text

Daniela COLTUC, Mihaela ILIE, Andreea MERLAN: A RLS model extracted by ICA
Abstract  Full Text

Adrian BURLACU, Cosmin COPOT, Corneliu LAZAR: SIFT based algorithm for point feature tracking
Abstract  Full Text

Octavian PASTRAVANU, Mihaela MATCOVSCHI: Invariance properties of time-varying linear systems
Abstract  Full Text

Adrian FILIPESCU, Alexandru STANCU, Silviu FILIPESCU: Sliding-mode adaptive control of Pioneer 3-DX wheeled mobile robot
Abstract  Full Text

Mohd Nazri ISMAIL, Abdullah Mohd ZIN: Impact of bandwidth for various services in heterogeneous network and quality of service via communication technologies in Malaysian higher educational
Abstract  Full Text

Nicolae PARASCHIV, Alexandru POPA, Cristina POPA: Online industrial process broadcast and control based on IP and serial networks
Abstract  Full Text

Iosif OLAH, Nicoleta Stefania HULEA, Lucian MASTACAN, Octavian Costel VORNICU: Thermal processes identification using virtual instrumentation
Abstract  Full Text

Traian MUNTEANU, Emil ROSU, Teodor DUMITRIU, Mihai CULEA, Romeo PADURARU, Cristinel DACHE: The optimal control for D.C. position drive system
Abstract  Full Text

Nicu BIZON: Buck supplies output voltage ripple reduction using fuzzy control
Abstract  Full Text

N. RABBAH, B. BENSASSI: Control of a multivariable web winding system
Abstract  Full Text

Nicu ROMAN, Bogdan IONITA, Dorel AIORDACHIOAE, Emil CEANGA, Ion BIVOL: Force prediction in cold rolling mills by polynomial methods
Abstract  Full Text

 

 

Cornelia NOVAC-UDUDEC, Rodica BIRLA: A Student Model in a multi-agent intelligent system


Abstract: The paper presents an intelligent tutoring system which provides a Student Model using the “overlay method”. The system’s tasks are distributed among the intelligent agents (software agents), each having clearly specified individual roles as following: administrator (Cerberus Agent); interface (InterfaceAgent); tutor (TutorAgent); student (StudentAgent). Moreover, the system is a “three-tier” architecture application and was implemented in Java language.

Keywords: multi-agent system, educational system, student model, intelligent system.

 

 

 

Nicolae JASCANU, Veronica JASCANU, Florin NICOLAU: A new approach to e-commerce multi-agent systems


Abstract: The negotiation process between two humans is an art. To capture the magic behind negotiation, a first step will be to build a model that incorporates human emotions and community emotional opinions. A model like this will be able to cope with the negotiation process easily, in a human-like way. The agents involved in a bilateral negotiation for some commodity should have some internal representation of human counterpart preferences. Using dimensional models of emotion, the agent is able to capture and interpret such knowledge. This paper settles the ground for emotional negotiations and the role of emotion in decision.

Keywords: negotiation, multi-agent systems, e-commerce, emotions, social shopping.

 

 

 

Adrian SERBENCU, Viorel MINZU, Adriana SERBENCU: An Ant Colony System based metaheuristic for solving single machine scheduling problem


Abstract: This paper proposes a metaheuristic for solving the Single Machine Scheduling Problem that is implemented by a hybrid system made up of an Ant Colony System and a stochastic descent algorithm called Kangaroo. The hybrid system is based on the collaboration between a social type multiagent system and an Iterated Solution Improvement method.

Keywords: metaheuristics, manufacturing, discrete optimization, stochastic descent, Ant Colony Systems.

 

 

 

Cornelia TUDORIE, Laurentiu FRANGU, Diana STEFANESCU: Automatic modeling of the linguistic values for database fuzzy querying


Abstract: In order to evaluate vague queries, each linguistic term is considered according to its fuzzy model. Usually, the linguistic terms are defined as fuzzy sets, during a classical knowledge acquisition off-line process. But they can also be automatically extracted from the actual content of the database, by an online process. In at least two situations, automatically modeling the linguistic values would be very useful: first, to simplify the knowledge engineer’s task by extracting the definitions from the database content; and second, where mandatory, to dynamically define the linguistic values in complex criteria queries evaluation. Procedures to automatically extract the fuzzy model of the linguistic values from the existing data are presented in this paper.

Keywords: Database, Flexible Query, Linguistic Values, Knowledge Acquisition.

 

 

 

Adina COCU: Knowledge-based diagnostic system with probabilistic approach


Abstract: This paper presents a knowledge learning diagnostic approach implemented in an educational system. Probabilistic inference is used here to diagnose knowledge understanding level and to reason about probable cause of learner’s misconceptions. When one learner takes an assessment, the system use probabilistic reasoning and will advice the learner about the most appropriate error cause and will also provide, the conforming part of theory which treats errors related to his misconceptions.

Keywords: diagnostic system, educational system, probability theory, graphical representation

 

 

 

Dan MUNTEANU: A quick survey of text categorization algorithms


Abstract: This paper contains an overview of basic formulations and approaches to text classification. This paper surveys the algorithms used in text categorization: handcrafted rules, decision trees, decision rules, on-line learning, linear classifier, Rocchio’s algorithm, k Nearest Neighbor (kNN), Support Vector Machines (SVM).

Keywords: information retrieval, algorithms, machine learning, text classification.

 

 

 

Dan MUNTEANU: Vector space model for document representation in information retrieval


Abstract: This paper presents the basics of information retrieval: the vector space model for document representation with Boolean and term weighted models, ranking methods based on the cosine factor and evaluation measures: recall, precision and combined measure.

Keywords: information retrieval, Boolean vector space model, term weighted vector space model, document representation.

 

 

 

Ana-Elena LUNGU, Daniela COLTUC, Laurentiu FRANGU: Computing negentropy based signatures for texture recognition


Abstract: The proposed method aims to provide a new tool for texture recognition. For this purpose, a set of texture samples are decomposed by using the FastICA algorithm and characterized by a negentropy based signature. In order to do recognition, the texture signatures are compared by means of Minkowski distance. The recognition rates, computed for a set of 320 texture samples, show a medium recognition accuracy and the method may be further improved.

Keywords: texture recognition, Independent Component Analysis, negentropy, Minkowski distance.

 

 

 

Daniela COLTUC, Mihaela ILIE, Andreea MERLAN: A RLS model extracted by ICA


Abstract: This paper presents some preliminary results regarding the ability of ICA in analyzing spectrometric data. A set of RLS spectra, measured for different concentration DNA solutions, in the presence of the probe TbDTPA, is decomposed like in a typical application of source separation and processed in order to get a model for RLS process in this case. The resulted curves fit well the model obtained by manually measuring and averaging the peaks of the RLS spectra.

Keywords: ICA, RLS spectra, DNA.

 

 

 

Adrian BURLACU, Cosmin COPOT, Corneliu LAZAR: SIFT based algorithm for point feature tracking


Abstract: In this paper a tracking algorithm for SIFT features in image sequences is developed. For each point feature extracted using SIFT algorithm a descriptor is computed using information from its neighborhood. Using an algorithm based on minimizing the distance between two descriptors tracking point features throughout image sequences is engaged. Experimental results, obtained from image sequences that capture scaling of different geometrical type object, reveal the performances of the tracking algorithm.

Keywords: SIFT feature, magnitude, orientation, descriptor, tracking.

 

 

 

Octavian PASTRAVANU, Mihaela MATCOVSCHI: Invariance properties of time-varying linear systems


Abstract: This paper formulates and proves two types of necessary and sufficient conditions for the characterization of positively (flow) invariant sets with respect to the state-space trajectories of the time-varying (non-autonomous) linear systems in both continuous- and discrete-time case. These conditions are expressed in terms of inequalities involving the matrix function that defines the system dynamics and a constant matrix that defines the shape of the invariant set. The first type of results refers to contractive invariant sets which decrease exponentially, and the second one considers invariant sets that remain constant. Our approach to non-autonomous systems accommodates, as particular cases, the elements of the invariant set analysis already elaborated for autonomous systems.

Keywords: time-invariant (autonomous) linear systems, time-varying (non-autonomous) linear systems, constant invariant sets, contractive invariant sets.

 

 

 

Adrian FILIPESCU, Alexandru STANCU, Silviu FILIPESCU: Sliding-mode adaptive control of Pioneer 3-DX wheeled mobile robot


Abstract: Parameter identification scheme and discrete-time adaptive sliding-mode controller applied to Pioneer 3-DX wheeled mobile robot (WMR) are presented in this paper. The dynamical model for mobile robot with one pair of active wheels, time–varying mass and moment of inertia have been used in sliding-mode control. Two closed-loop, on-line parameter estimators have been used in order to achieve robustness against parameter uncertainties (robot mass and moment of inertia). Two sliding-mode adaptive controllers corresponding to angular and position motion have been designed. Closed-loop circular trajectory tracking Pioneer 3-DX real-time control is presented.

Keywords: Discrete-time Pioneer 3-DX model, sliding-mode adaptive control, on-line parameter estimation.

 

 

 

Mohd Nazri ISMAIL, Abdullah Mohd ZIN: Impact of bandwidth for various services in heterogeneous network and quality of service via communication technologies in Malaysian higher educational


Abstract: This study focuses on heterogeneous services and communication technologies for retrieving and sending information between students and lecturers in Malaysian higher educational institutes. It intends to investigate pattern and significant level of various services implementation, convergence of communication technologies and bandwidth capacity for last mile users (students and lecturers). It is designed to determine which communication technologies, services, bandwidth capacity and QoS will achieve the highest level acceptance in Malaysian higher educational institutes. Heterogeneous environment can also generate mobility approaches. A survey method is used to collect data from respondents (System Network Administrator) in Malaysian higher educational institutes. Statistical Analysis using t-test shows that implementation of high speed bandwidth for Internet (WAN) achieved significant level. Frequency test are used to analyze the various services implementation via convergence of communication technologies in heterogeneous environment for retrieving information. The most apparent finding of this study is that various services implementation in higher educational institutes can affect convergence of communication technologies usage in accessing information over heterogeneous network environment between students and lecturers. The study or analysis confirms that Malaysia higher educational institutes have not achieved a maximum level of various services implementation via convergence of communication technologies between students and lecturers in heterogeneous network environment. Therefore, low convergence of communication technologies can produce a low mobility and ubiquitous interaction between students and lectures in Malaysian higher educational institutes.

Keywords: LAN, Heterogeneous Environment, QoS, Ubiquitous, Mobility, WAN.

 

 

 

Nicolae PARASCHIV, Alexandru POPA, Cristina POPA: Online industrial process broadcast and control based on IP and serial networks


Abstract: The Internet network in conjunction with a strong security policy can be an ideal online broadcast system for the industrial processes monitoring and control. Because the distributed controllers are connected through industrial serial connection, the network between controllers cannot be extended very much. The problem appears when the plant is away and cannot be accessed via a serial connection. For administrative and control purposes the serial network can be interconnected with an IP network. In this case the operator can be anywhere in the world and can interact via an interface with the process. This paper will describe flow process remote controlled using a mixture of two communication technologies: RS485 and TCP/IP. The junction between these technologies is transparent and the operator is able to monitor the process using an interface console.

Keywords: control, fieldbus, industrial network, Ethernet, serial-communication, security.

 

 

 

Iosif OLAH, Nicoleta Stefania HULEA, Lucian MASTACAN, Octavian Costel VORNICU: Thermal processes identification using virtual instrumentation


Abstract: In this paper the experimental identification problem of thermal processes is presented, in order to establish their mathematical models which permit the adoption of the automation solutions, respectively the specification of a suitable control law. With this aim in view, the authors resorted to use Virtual Instrumentation with the aid of the LabVIEW development medium. In order to solve the problem of acquisition and processing data from physical real processes, Virtual Instruments which provide at the end a mathematical model which is basis of choosing the automation equipment of the aim followed was designed and achieved. The achieved Virtual Instruments get the opportunity to be used either in student instruction field with the virtual processes identification techniques or to put the identification of some real processes to good use of diverse beneficiaries. The results of some experimental attempts which were achieved during different thermal processes, illustrate the utility of the demarches performed in this paper.

Keywords: experimental identification, identification, industrial processes, process, thermal process, virtual instrumentation.

 

 

 

Traian MUNTEANU, Emil ROSU, Teodor DUMITRIU, Mihai CULEA, Romeo PADURARU, Cristinel DACHE: The optimal control for D.C. position drive system


Abstract: A position drive system with D.C. motor is operated using conventional position, speed and current controls. These controls provide a good dynamic and steady-state behaviour, but they do not take in consideration the conversion’s efficiency. It is very well known that in the transient behaviour, as starting and stopping, the conversion efficiency is diminished to the value about 50 per cent, while in the steady-state it is greater. The main goal of the paper is to develop a new optimal type control law, which minimizes the drawn energy used for the covering of a given position trajectory. The synthesis of the optimal energetic control law is accomplished by comparing the two control methods, conventional and optimal. The experimental results, via the simulation procedure, are also presented.

Keywords: optimal control, energetic criteria, D.C. position drive system.

 

 

 

Nicu BIZON: Buck supplies output voltage ripple reduction using fuzzy control


Abstract: Using the PWM control for switching power supplies the peaks EMI noise appear at the switching frequency and its harmonics. Using randomize or chaotic PWM control techniques in these systems the power spectrum is spread out in all frequencies band spectral emissions, but with a bigger ripple in the output voltage. The proposed nonlinear feedback control method, which induces chaos, is based by fuzzy rules that minimize the output voltage ripple. The feasibility and effectiveness of this relative simple method is shown by simulation. A comparison with the previous control method is included, too.

Keywords: fuzzy control, anticontrol of chaos; buck converter; power spectrum; output voltage ripple.

 

 

 

N. RABBAH, B. BENSASSI: Control of a multivariable web winding system


Abstract: To guarantee a good productivity and a better quality of the industrial systems, it is necessary to know the evolution of its various parameters but also its operation mode. The main goal of this work is to carry out a study to improve a flatness control in a reversing cold rolling mill, precisely the web winding system part. We designate by a winding process any system applying the cycles of unwinding, transport, treatment and winding to various flat products. This system knows several constraints such as the thermal effects caused by the frictions, and the mechanical effects provoked by metal elongation, that generates dysfunctions due to the influence of the process conditions. For this installations type, the various automatisms functions, often very advanced, are realized in modular systems with distributed architecture. Our main goal is to obtain a precise thickness, with the best possible regularity. With this intention, we are preceded to the modelling and the control of the nonlinear dynamic behaviour of a web winding process.

Keywords: Web winding system (WWS), Modelling, Nonlinear Control, Linearizing Control, Diffeomorphism.

 

 

 

Nicu ROMAN, Bogdan IONITA, Dorel AIORDACHIOAE, Emil CEANGA, Ion BIVOL: Force prediction in cold rolling mills by polynomial methods


Abstract: A method for steel and aluminium strip thickness control is provided including a new technique for predictive rolling force estimation method by statistic model based on polynomial techniques.

Keywords: Cold rolling, Predictive model, Feed-forward control, Statistic Polynomial Methods.