OPTIMIZATION MODEL AND DSS FOR MAXIMUM RESOLUTION DICHOTOMIES

James K. Ho, Sydney C. K. Chu and S. S. Lam

A topological model is presented for complex data sets in which the attributes can be cast into a dichotomy. It is shown that the relative dominance of the two parts in such a dichotomy can be measured by the corresponding areas in its star plot. An optimization model is proposed to maximize the resolution of such a measure by choice of configuration of the attributes, as well as the angles among them. The approach is illustrated with the case of online auction markets, where there is a buyer-seller dichotomy as to whether conditions are favourable to buyers or sellers. An implementation of the methodology in a spreadsheet based DSS is demonstrated. Its ease of use is promising for diverse applications.

IMITATING THE KNOWLEDGE MANAGEMENT OF COMMUNITIES OF PRACTICE

Juan Pablo Soto, Aurora Vizcaíno, Javier Portillo, Oscar M. Rodríguez-Elias and Mario Piattini

Advances in technology have led to the development of knowledge management systems with the intention of improving organizational performance. Nevertheless, implementation of this kind of mechanisms is not an easy task due to the necessity of taking into account social aspects (such as reputation) that improve the exchange of information between groups of people. Considering, the advantages of working with groups with similar interests we have modelled communities of agents which represent communities of people interested in similar topics. In order to implement this model we propose a multi-agent architecture in charge of evaluating the relevance of the knowledge in a knowledge base and the degree of reputation that a person has as the contributor of information. We pay particular attention to showing how the use of the agents works by using a prototype system to search for knowledge related to a particular domain of a community of practice. Several communities of agents integrated into an organization have the capacity to follow the interaction process of employees when carrying out their daily activities.

NONLINEAR FUZZY SELFTUNING PID CONTROL TECHNOLOGY AND ITS APPLICATIONS IN AUTOMATED PROGRAMMING ROBOTICS

Ganwen Zeng and Qianglong Zeng

The paper presents an advanced Fuzzy self-tuning PID controller theory and it implement its applications on Data I/O’s automated robotic programming systems. Considerable programming technology shift occurred in recent device programmer industry; programming density have been constantly fast growing from low-volume to high-volume programming for all kinds of non-volatile flash memory devices such as NOR flash, NAND flash, and MMC cards, SD flash cards, serial flash device, serial flash cards, flash-based microcontrollers and flash disks as high performance M-systems DiskOnChip. Device programming mode is more demanding an automatic programming than manual operation mode. It drives the creation and implementation of a high-performance automated programming robotic systems. This paper shows how this proposed advanced Fuzzy self-tuning PID controllers work on these automated programming robotic automation systems.

MULTIVARIATE CONTROL CHARTS WITH A BAYESIAN NETWORK

Sylvain Verron, Teodor Tiplica and Abdessamad Kobi

The purpose of this article is to present an approach allowing the fault detection of a multivariate process with a bayesian network. As a discriminant analysis is easily modeled with a bayesian network, we will show that we we can consider the multivariate T2 and MEWMA control charts as particular cases of the discriminant analysis. So, we give the structure of the bayesian network as well as the parameters of the network in order to detect faults in the multivariate space in the same manners as if we used multivariate control charts. The resulting bayesian network, with a computed threshold, is similar to the multivariate control charts.

OBJECT LIST CONTROLLED PROCESS DATA SYSTEM

Anton Scheibelmasser and Bernd Eichberger

The appropriate design of a system is one of the essential topics at the beginning of a new development project. According to the intended purpose of a device the first step is to model the system in order to get a structure for the implementation of the required features. In general the implementation of the system requirements is split in hardware parts and tasks which are done in software. In case of the hardware design the solutions for the challenges are mostly clear and supported by fundamentals of e.g. digital logic laws and several design methods. If we think of the software part a lot of problems have to be solved without such clear fundamentals. Object oriented design is one of the paradigms which promise a way for designing stable and reliable software. A problem arises in this context if the used microprocessor platform is not supported with a compiler for an object oriented programming language. In this case only the system modelling could be done in terms of software objects and their relations, the implementation has to be done in a procedural language. The following article is based on research work done in the development of a modular process data system. Based on a sequential main program and interrupt driven hardware interfaces, a software implementation without an operating system was implemented. By means of special software structure called linked object list, object oriented design was implemented with the procedural language “C”. Due to this design a reusable and flexible system was achieved which enables a high degree of flexibility concerning the hardware configuration and system customization at the user site.

DETECTION OF THE NEED FOR A MODEL UPDATE IN STEEL MANUFACTURING

Heli Koskimäki (née Junno), Ilmari Juutilainen, Perttu Laurinen and Juha Röning

When new data are obtained or simply when time goes by, the prediction accuracy of models in use may decrease. However, the question is when prediction accuracy has dropped to a level where the model can be considered out of date and in need of updating. This article describes a method that was developed for detecting the need for a model update. The method is applied in the steel industry, and the model whose need of updating is under study is a regression model developed to model the yield strength of steel plates. It is used to plan process settings in steel plate product manufacturing. To decide on the need for updating, information from similar past cases was utilized by introducing a limit called an exception limit. The limit was used to indicate when a new observation was from an area of the model input space where the prediction errors of the model have been too high. Moreover, an additional limit was formed to indicate when too many exceedings of the exception limit have occurred within a certain time scale. These two limits were then used to decide when to update the model.

DIGITAL PATTERN SEARCH AND ITS HYBRIDIZATION WITH GENETIC ALGORITHMS FOR GLOBAL OPTIMIZATION

Nam-Geun Kim, Youngsu Park and Sang Woo Kim

In this paper, we present a new approach of evolutionary algorithm called genetic pattern search algorithm (GPSA). The proposed algorithm is closely related to genetic algorithms which use binary-coded genes. The main contribution of this paper is to propose binary-coded pattern called digital pattern which is transformed from real-coded pattern in general pattern search methods. Moreover we offer self-adapting genetic algorithm by adopting digital pattern that modifies the step size and encoding esolutions of previous optimization procedure, and chases the optimal pattern's direction. In addition, we compare GPSA with genetic algorithm in the robustness and the performances of optimization. All experiments employ the well-known benchmark functions whose functional values and coordinates of each global minimum are already reported.

DISTRIBUTED EMBEDDED SYSTEM FOR ULTRALIGHT AIRPLANE MONITORING

J. Kotzian and V. Srovnal Jr.

This paper presents distributed embedded monitoring system that is developed for small aircrafts, sports airplane and ultralights airplanes. System is made from modules connected by industrial bus CAN. This low cost system is trying to solve bad situation with many ultralights without any digital measurement unit due to their prices. The contribution shows basic architecture of the embedded monitoring system and presents some parts of hardware and software implementation. The interface between aviator and airplane is established using graphic user interface based on operating system uClinux.

DISCRETE DYNAMIC SLIDING SURFACE CONTROL FOR ROBUST SPEED CONTROL OF INDUCTION MACHINE DRIVE

Abdel Faqir, Daniel Pinchon, Rafiou Ramanou and Sofiane Mahieddine

This paper proposes the discrete dynamic sliding surface control to guarantee the existence of discrete sliding mode and reduce the chattering phenomena for speed control of induction machine drive. In discrete systems, the controller does not control the system during the sampling interval. The great chattering and large control signal are caused by the high switching gain. In this paper, the dynamic sliding surface is introduced to overcome the drawback. By setting the initial value of the dynamic sliding surface, the system can lock to the sliding surface quickly without high switching gain. The control signal can be reduced and the chattering can be eliminated. Furthermore, the induction machine speed control system is used to show this controller’s robustness to against the parameter variation and external load. The speed of the induction machine is regulated using the indirect field oriented control (IFOC). Thus, after the application of the IFOC technique by determining the decoupled model of the machine, a discrete sliding surface controller has been applied. Simulation study is used to show the performances of the proposed method and then validated by an experimental prototype.

BOUNDARY CONTROL OF A CHANNEL - Last Improvements

Valérie dos Santos and Christophe Prieur

Different improvements have been developed in regards to the stability and the control of two-by-two non linear systems of conservation laws, and in particular for the Saint-Venant equations and the control of flow and water level on irrigation channel. One stability result based on the Riemann coordinates is presented here and sufficient conditions are given to insure the Cauchy convergence. Another result still based on the Riemann approach is presented too, in the linear case, to improve the feedback control based on the Riemann invariants.

EFFECTIVE GENETIC OPERATORS OF COOPERATIVE GENETIC ALGORITHM FOR NURSE SCHEDULING

Makoto Ohki, Shin-ya Uneme, Shigeto Hayashi and Masaaki Ohkita

This paper proposes effective genetic operators for cooperative genetic algorithm (GA) to solve a nurse scheduling problem. In hospitals, 15-30 nurses are assigned to any section such as the internal medicine department or the pediatrics department. A clinical director of the department makes a duty schedule of all nurses of the department every month. Such the scheduling is very complex task. It takes one or two weeks to create the nurse schedule even by a veteran director. Recently, computer software creating the nurse schedule is developed to reduce such the problem. Even when the newest commercial software creates and optimizes the nurse schedule, it needs more than one or two hours. Since this is very risky for the users, an algorithm giving a solution in a shorter running time is still required. In conventional ways using the cooperative GA, a crossover operator is only employed for the optimization, because it does not lose consistency between chromosomes. We propose mutation operator and virus operator for the cooperative GA, which does not lose consistency of the nurse schedule. The cooperative GA with these new operators has brought a surprisingly good result, it has never been brought by the conventional algorithm.

BINARY OPTIMIZATION: A RELATION BETWEEN THE DEPTH OF A LOCAL MINIMUM AND THE PROBABILITY OF ITS DETECTION

B. V. Kryzhanovsky, V. M. Kryzhanovsky and A. L. Mikaelian

The standard method in optimization problems consists in a random search of the global minimum: a neuron network relaxes in the nearest local minimum from some randomly chosen initial configuration. This procedure is to be repeated many times in order to find as deep an energy minimum as possible. However the question about the reasonable number of such random starts and whether the result of the search can be treated as successful remains always open. In this paper by analyzing the generalized Hopfield model we obtain expressions describing the relationship between the depth of a local minimum and the size of the basin of attraction. Based on this, we present the probability of finding a local minimum as a function of the depth of the minimum. Such a relation can be used in optimization applications: it allows one, basing on a series of already found minima, to estimate the probability of finding a deeper minimum, and to decide in favor of or against further running the program. The theory is in a good agreement with experimental results.

A NEW LOAD ADJUSTMENT APPROACH FOR JOB-SHOPS

Z. Bahroun, J.-P. Campagne and M. Moalla

This paper presents a new load adjustment approach by overlapping for a set of jobs in a job-shop context, guaranteeing the existence of a limited capacity schedule without scheduling under the assumption of pre-emptive tasks. This approach is based on the exploitation of the tasks scheduling time segments overlapping and on the distribution of the job’s margins between tasks in a just in time context. First, we present a literature review concerning load adjustment approaches. Second, we introduce the overlapping load adjustment approach. Third, we present an original heuristic to use this approach in the case of job-shops organized firms. After that, we present the scheduling approach. Finally, we will discuss a more general use of this approach and the possible extensions.

SATURATION FAULT-TOLERANT CONTROL FOR LINEAR PARAMETER VARYING SYSTEMS

Ali Abdullah

This paper presents a methodology for designing a fault-tolerant control (FTC) system for linear parameter varying (LPV) systems subject to actuator saturation fault. The FTC system is designed using linear matrix inequality (LMI) and model estimation techniques. The FTC system consists of a nominal control, fault diagnostic, and fault accommodation schemes. These schemes are designed to achieve stability and tracking requirements, estimate a fault, and reduce the fault effect on the system. Simulation studies are used to illustrate the proposed design.

LINEAR PROGRAMMING FOR DATABASE ENVIRONMENT

Akira Kawaguchi and Jose Alfredo Perez

Solving large-scale optimization problems requires an integration of data-analysis and data-manipulation capabilities. Today's databases support real-time decision analysis and complex decision making. Nevertheless, little attempt has been made to facilitate general linear programming solvers for database environments. Dozens of sophisticated tools and software libraries that implement linear programming model can be found. But, there is no database-embedded linear programming tool seamlessly and transparently utilized for database processing. The focus of this study is to fill out this kind of technical gap of data analysis and data manipulation, in the event of solving large-scale linear programming problems for the applications built on the database environment. Specifically, this paper studies the representation of the linear programming model in relational structures and the computational method to solve the linear programming problems. This development is critical in the circumstances of the wide applicability of the linear programming problems to today's database applications. Foundations for and preliminary experimental results of this study are presented.

BREAKING ACCESSIBILITY BARRIERS - Computational Intelligence in Music Processing for Blind People

Wladyslaw Homenda

A discussion on involvement of knowledge based methods in implementation of user friendly computer programs for disabled people is the goal of this paper. The paper presents a concept of a computer program that is aimed to aid blind people dealing with music and music notation. The concept is solely based on computational intelligence methods involved in implementation of the computer program. The program is build around two research fields: information acquisition and knowledge representation and processing which are still research and technology challenges. Information acquisition module is used for recognizing printed music notation and storing acquired information in computer memory. This module is a kind of the paper-to-memory data flow technology. Acquired music information stored in computer memory is then subjected to mining implicit relations between music data, to creating a space of music information and then to manipulating music information. Storing and manipulating music information is firmly based on knowledge processing methods. The program described in this paper involves techniques of pattern recognition and knowledge representation as well as contemporary programming technologies. It is designed for blind people: music teachers, students, hobbyists, musicians.

MULTICRITERIA DECISION MAKING IN BALANCED MODEL OF FUZZY SETS

Wladyslaw Homenda

In the paper aspects of negative information and of information symmetry in context of uncertain information processing is considered. Both aspects are presented in frames of fuzzy sets theory involved in data aggregation and decision making process. Asymmetry of classical fuzziness and its orientation to positive information are pointed out. The direct dependence of symmetry of uncertain information on negative information maintenance is indicated. The symmetrical, so called balanced, extension of classical fuzzy sets integrating positive and negative information an paralleling positiveness/negativeness with symmetry of fuzziness is presented. Balanced counterparts of classical fuzzy connectives are introduced.

MORE EXPRESSIVE PLANNING GRAPH EXTENSION

Joseph Zalaket and Guy Camilleri

Since its appearance Graphplan allured the researchers in AI planning for its compact structure. In addition to its performance to solve planning problems, Graphplan has served many heuristic planners by its planning graph structure. Many extensions have been made to the Graphplan or to its planning graph to enhance their performance and to make them able to solve new type of knowledge like temporal and resources. The most of these extensions have treated the temporal and numeric resource knowledge as a foreign body incorporated into Graphplan. Our deep observation to the Graphplan structure showed us that this structure is able to deal with all kind of knowledge by the same way as with symbolic knowledge. Even more, this structure is able to handle black box functions which manipulate all kind of data. In this paper, we present a variation of Graphplan which supports the execution of external functions for numeric knowledge update. This variation allows Graphplan to run all kind of knowledge using its original planning graph as the base of the data structure.

NONLINEAR MODEL PREDICTIVE CONTROL OF A LINEAR AXIS BASED ON PNEUMATIC MUSCLES

Harald Aschemann and Dominik Schindele

This paper presents a nonlinear optimal control scheme for a mechatronic system consisting of a guided carriage driven by an antagonistic pair of pneumatic muscle actuators. Modelling leads to a system of nonlinear differential equations including polynomial approximations of the volume characteristic as well as the force characteristic of the pneumatic muscles. The proposed control has a cascade structure. The nonlinear norm-optimal control of both pneumatic muscle pressures is based on an approximative solution of the corresponding HJB-equation, whereas the outer control loop involves a multivariable NMPC of the carriage position and the mean internal pressure of the pneumatic muscles. To improve the tracking behaviour, the feedback control loops are extended with nonlinear feedforward control based on differential flatness. Remaining model uncertainties as well as nonlinear friction can be counteracted by an observer-based disturbance compensation. Experimental results from an implementation on a test rig show an excellent control performance.

GENETIC REINFORCEMENT LEARNING OF FUZZY INFERENCE SYSTEM APPLICATION TO MOBILE ROBOTIC

Abdelkrim Nemra, Hacene Rezine and Abdelkrim Souici

An efficient genetic reinforcement learning algorithm for designing Fuzzy Inference System (FIS) with out any priory knowledge is proposed in this paper. Reinforcement learning using Fuzzy Q-Learning (FQL) is applied to select the consequent action values of a fuzzy inference system, in this method, the consequent value is selected from a predefined value set which is kept unchanged during learning and if the optimal solution is not present in the randomly generated set, then the performance may be poor. Also genetic algorithms (Genetic Algorithm) are performed to on line search for better consequent and premises parameters based on the learned Q-values as adaptation function. In Fuzzy-Q-Learning Genetic Algorithm (FQLGA), memberships (premises) parameters are distributed equidistant and the consequent parts of fuzzy rules are randomly generated. The algorithm is validated in simulation and experimentation on mobile robot reactive navigation behaviors.

DEFECT-RELATED KNOWLEDGE ACQUISITION FOR DECISION SUPPORT SYSTEMS IN ELECTRONICS ASSEMBLY

Sébastien Gebus and Kauko Leiviskä

Real-time process control and production optimization are extremely challenging areas. Traditional approaches often do not work due to a lack of robustness or reliability, especially when dealing with incomplete, inaccurate, or simply irrelevant data. This is a major problem when building decision support systems especially in electronics manufacturing, where it is quite common to have large databases and run blindly feature extraction and data mining methods. Performance of these methods could, however, be drastically increased when combined with knowledge or expertise of the process. This paper describes how defect-related knowledge on an electronic assembly line can be integrated in the decision making process at an operational and organizational level. It focuses in particular on the efficient acquisition of shallow knowledge concerning everyday human interventions on the production lines, as well as on the conceptualization and factory wide sharing of the resulting defect information. Software with dedicated interfaces has been developed for that purpose. Semi-automatic knowledge acquisition from the production floor and generation of comprehensive reports for the quality department resulted in an improvement of the usability, usage, and usefulness of the decision support system.

A NEURAL-CONTROL SYSTEM FOR A HUMANOID ARTIFICIAL ARM

Michele Folgheraiter, Giuseppina Gini and Massimo Cavallari

In this paper we illustrate the architecture and the main features of a bio-inspired control system employed to govern an anthropomorphic artificial Arm. The manipulation system we developed was designed starting from an attentive study of the human limb from the anatomical, physiological and neurological point of view. In accordance with the general view of the Biorobotics field we try to replicate the structure and the functionalities of the natural limb. Thanks to this biomimetic approach we obtained a system that can perform movements similar to those of the natural limb. The control system is organized in a hierarchical way. The low level controller emulates the neural circuits located in the human spinal cord and is charged to reproduce the reflexes behaviors and to control the arm stiffness. The high level control system generates the arm trajectory performing the inverse kinematics and furnishing the instantaneous muscles reference position. In particular we implemented the Inverse kinematic using a gradient based algorithm; at each step the actuators movements are arranged in order to decrease the distance between the wrist and the target position. Simulation and experimental results shows the ability of the control system in governing the arm to follow a predefined trajectory and to perform human like reflexes behaviors.

COMPARYING A TABU SEARCH PROCESS - Using and Not Using and Intensification Strategy to Solve the Vehicle Routing Problem

Etiene Pozzobom Lazzeris Simas and Arthur Tórgo Gómez

In this paper we propose a Tabu Search algorithm to solve the Vehicle Routing Problem. The Vehicle Routing Problem are usually defined as the problem that concerns in creation of least cost routs to serve a set of clients by a fleet of vehicles. We develop an intensifications strategy to diversify the neighbours generated and to increase the neighbourhood size. We had done experiments using and not using the intensification strategy to compare the performance of the search. The experiments we had done showed that an intensification strategy allow an increase on the solutions quality.

A DISCRETE-EVENT SYSTEM APPROACH TO MULTI-AGENT DISTRIBUTED CONTROL OF CONTAINER TERMINALS

Guido Maione

The area of managing and controlling intermodal terminal systems is relatively new. The paradigms of Discrete Event Systems for modelling purpose and of Multi-Agent Systems for distributed control seem promising. Many research attempts have been made to develop modelling and simulation tools but no standard exists. This paper presents a Discrete Event System model of the agents introduced to describe how a distributed control of the terminal activities can be achieved. The interaction mechanism between four classes of agents is modelled in detail. The approach is useful to develop a simulation platform to test MAS efficiency in terminal management and to measure the performance of static or adapted control strategies.

A DISTRIBUTED REINFORCEMENT LEARNING CONTROL ARCHITECTURE FOR MULTI-LINK ROBOTS - Experimental Validation

Jose Antonio Martin H. and Javier De Lope

A distributed approach to Reinforcement Learning (RL) in multi-link robot control tasks is presented. One of the main drawbacks of classical reinforcement learning is the combinatorial explosion when multiple states variables and multiple actuators are needed to optimally control a complex agent in a dynamical environment. In this paper we present an approach to avoid this drawback based on a distributed RL architecture. The experimental results in learning a control policy for diverse kind of multi-link robotic models clearly shows that it is not necessary that each individual RL-agent perceives the complete state space in order to learn a good global policy but only a reduced state space directly related to its own environmental experience. The proposed architecture combined with the use of continuous reward functions results of an impressive improvement of the learning speed making tractable some learning problems in which a classical reinforcement learning with discrete rewards (-1,0,1) does not work.

ROBUST ADAPTIVE WAVELET NEURAL NETWORK TO CONTROL A CLASS OF NONLINEAR SYSTEMS

A. Hussain, N. Essounbouli, A. Hamzaoui and J. Zaytoon

This paper deals with the synthesis of a Wavelet Neural Network adaptive controller for a class of second order systems. Due to its fast convergence, the wavelet neural network is used to approximate the unknown dynamics, which will be on-line adjusted according to the adaptation laws deduced from the stability analysis. To ensure the robustness of the closed loop system, a modified sliding mode control signal is used. In this work, variable sliding surface is considered to reduce the starting energy without deteriorating the tracking performances. Furthermore, the knowledge of the upper bounds of both the external disturbances and the approximation errors is not needed. The global stability of the closed loop system is guaranteed in the sense of Lyapunov. Finally, a simulation example is presented to illustrate the efficiency of the

PATTERN-DRIVEN REUSE OF EMBEDDED CONTROL DESIGN - Behavioral and Architectural Specifications in Embedded Control System Designs

Miroslav Sveda, Ondrej Rysavy and Radimir Vrba

This paper deals with reuse of architectural and behavioral specifications of embedded systems employing finite-state and timed automata. The contribution proposes not only how to represent a system’s formal specification as an application pattern structure of specification fragments, but also how to measure similarity of formal specifications for retrieval with case-based reasoning support. The paper provides also an insight into case-based reasoning support as applied to formal specification reuse by application patterns built on finite-state and timed automata. Those application patterns create a base for a pattern language supporting reuse-oriented design process for a class of real-time embedded systems.

A SERVICE-ORIENTED FRAMEWORK FOR MANNED AND UNMANNED SYSTEMS TO SUPPORT NETWORK-CENTRIC OPERATIONS

Norbert Oswald, André Windisch, Stefan Förster, Herwig Moser and Toni Reichelt

Network-centricity and autonomy are two buzzwords that have found increasing attention since the beginning of this decade in both, the military and civil domain. Although various conceptions exist of which capabilities are required for a system to be considered network-centric or autonomous, there can hardly be found proposals or prototypes that describe concrete transformations for both capabilities into software. The presented paper reviews work accomplished at EADS Military Air Systems driven by the need to develop an infrastructure that supports the realisation of both concepts in software with respect to traditional and modern software engineering principles, e.g., re-use and service-oriented development. This infrastructure is provided in form of a prototypical framework, accompanied by configuration and monitoring tools. Tests in a complex scenario requiring network-centricity and autonomy have shown that a significant technical readiness level can be reached by using the framework for mission software development.

A FUZZY PARAMETRIC APPROACH FOR THE MODEL-BASED DIAGNOSIS

F. Lafont, N. Pessel and J. F. Balmat

This paper presents a new approach for the model-based diagnosis. The model is based on an adaptation with a variable forgetting factor. The variation of this factor is managed thanks to fuzzy logic. Thus, we propose a design method of a diagnosis system for the sensors defaults. In this study, the adaptive model is developed theoretically for the Multiple-Input Multiple-Output (MIMO) systems. We present the design stages of the fuzzy adaptive model and we give details of the Fault Detection and Isolation (FDI) principle. This approach is validated with a benchmark: an hydraulic process with three tanks. Different defaults (sensors) are simulated with the fuzzy adaptive model and the fuzzy approach for the diagnosis is compared with the residues method. The first results obtained are promising and seems applicable on a set of MIMO systems.

TRACKING CONTROL DESIGN FOR A CLASS OF AFFINE MIMO TAKAGI-SUGENO MODELS

Carlos Ariño, Antonio Sala and Jose Luis Navarro

When controlling Takagi-Sugeno fuzzy systems, verification of some sector conditions is usually assumed. However, setpoint changes may alter the sector bounds. Alternatively, setpoint changes may be considered as an offset addition in many cases, and hence affine Takagi-Sugeno models may be better suited to this problem. This work discusses a nonconstant change of variable in order to carry out offset-ellimination in a class of MIMO canonical affine Takagi-Sugeno models. Once the offset is cancelled, standard fuzzy control design techniques can be applied for arbitrary setpoints. The canonical models studied use as state representation a set of basic variables and their derivatives. Some examples are included to illustrate the procedure.

BEHAVIOUR NAVIGATION LEARNINIG USING FACL ALGORITHM

Abdelkarim Souissi and Hacene Rezine

In this article, we are interested in the reactive behaviours navigation training of a mobile robot in an unknown environment. The control consists in bringing the robot in a given position, avoiding obstacles and releasing it from the tight corners and deadlock obstacles shape. In this framework, we used the reinforcement learning (FACL) method, or Fuzzy Actor-Critic learning based on temporal differences prediction method (TD). It allows the output adaptation of fuzzy inference system apprentice (SIF) in response to the rewards and punishments which it receives when interacting with the environment. The system has continuous type states and actions.

A JOINT HIERARCHICAL FUZZY-MULTIAGENT MODEL DEALING WITH ROUTE CHOICE PROBLEM - RoSFuzMAS

Habib M. Kammoun, Ilhem Kallel and Adel M. Alimi

Nowadays, multiagent architectures and traffic simulation agent-based are the most promising strategies for intelligent transportation systems. This paper presents a road supervision model based on fuzzy-multiagent system and simulation, called RoSFuzMAS. Thanks to agentification of all components of the transportation system, dynamic agents interact to provide real time information and a preliminary choice of advised routes. To ensure the model rationality, and to improve the route choice make decision, we propose to use a hierarchical Fuzzy inference including some pertinent criteria handling the environment as well as the driver behavior. A multiagent simulator with graphic interface has been achieved to visualize, test and discuss our road supervision system. Experimental results demonstrate the capability of RoSFuzMAS to perform a dynamic path choice minimizing traffic jam occurrences by combining multiagent technology and real time fuzzy behaviors.

TARGET VALUE PREDICTION FOR ONLINE OPTIMIZATION AT ENGINE TEST BEDS

Alexander Sung, Andreas Zell, Florian Kl¨opper, Alexander Vogel

The settling times of target functions play an important role in the domain of online optimization at the engine test bed. Inert target functions generally induce long measuring times which lead to increased costs. In this article, we analyze how previous knowledge about the physical behavior of target functions can be used to early predict the final steady state value to reduce measuring times.

DISCRETE GENETIC ALGORITHM AND REAL ANT COLONY OPTIMIZATION FOR THE UNIT COMMITMENT PROBLEM

Guillaume Sandou

In this paper, a new cooperative metaheuristic for the solution of the classical Unit Commitment problem is presented. This problem is known to be an often large scale, mixed integer programming problem. Due to the curse of combinatorial complexity, the exact solution is often intractable. Thus, a metaheuristic based method has to be used to compute a very often suitable solution with low computation times. A new approach is presented here. The main idea is to couple a genetic algorithm to compute binary variables (on/off status of units), and an ant colony based algorithm to compute real variables (produced powers). Finally, results show that the cooperative method leads to the tractable computation of a satisfying solution for the Unit Commitment problem.

NEW RESULTS ON DIAGNOSIS BY FUZZY PATTERN RECOGNITION

Mohamed Saïd Bouguelid, Moamar Sayed Mouchaweh and Patrice Billaudel

We use the classification method Fuzzy Pattern Matching (FPM) to realize the industrial and medical diagnosis. FPM is marginal, i.e., its global decision is based on the selection of one of the intermediate decisions. Each intermediate decision is based on one attribute. Thus, FPM does not take into account the correlation between attributes. Additionally, FPM considers the shape of classes as convex one. These drawbacks make FPM unusable for many real world applications. In this paper, we propose to improve FPM to solve these drawbacks. Several synthetic and real data sets are used to show the performances of the Improved FPM (IFPM) with respect to classical one as well as to the well known classification method K Nearest Neighbours (KNN). KNN is known to be preferment in the case of data represented by correlated attributes or by classes with non convex shape.

INVERSION OF A SEMI-PHYSICAL ODE MODEL

Laurent Bourgois, Gilles Roussel and Mohammed Benjelloun

This study proposes to examine the performances of an inverse dynamic model by fusion of statistical training and deterministic modeling. We carry out an inverse semi-physic model using a recurrent neural network. The structure of this network is guided by preliminary search of a reverse discrete state form of the direct model. The performances in term of generalization, regularization and training effort are highlighted compared to the reduction in parameters to estimate of the neural network. Some tests are carried out on a simple second order model, but the form of a dynamic system characterized by an ordinary differential equation of an unspecified $r$ order is proposed.

TAKAGI-SUGENO MULTIPLE-MODEL CONTROLLER FOR A CONTINUOUS BAKING YEAST FERMENTATION PROCESS

Enrique Herrera, Bernardino Castillo, Jesús Ramírez and Eugénio C. Ferreira

The purpose of this work is to design a fuzzy integral controller to force the switching of a bioprocess between two different metabolic states. A continuous baker’s yeast culture is divided in two sub-models: a respiro-fermentative with ethanol production and a respirative with ethanol consumption. The switching between both different metabolic states is achieved by means of tracking a reference substrate signal. A substrate fuzzy integral controller model using sector nonlinearity was built for both nonlinear models; the controller gains were designed using Linear Matrix Inequalities (LMI’s).

TOWARDS RELIABLE AUTOFOCUSING IN AUTOMATED MICROSCOPY

Silvie Luisa Brázdilová

The results presented in this paper are twofold. First, autofocusing in automated microscopy is studied and evaluated with respect to biomedical samples whose images can have more focal planes. While the proposed procedure for finding the maximum of a focus function in a short time works satisfactorily, the focus function itself is identified as the weakest link of the whole process. Second, an interesting property of functions used for genetic programming, and an algorithm for generating new individuals are introduced. Their usefulness and applicability are demonstrated on the problem of finding a new focus function for automated autofocusing in microscopy.

A HYBRID INTELLIGENT MULTI-AGENT METHOD FOR MONITORING AND FAULTS DIAGNOSIS

Gang Yao and Tianhao Tang

This paper presents a hybrid intelligent multi-agent method for monitoring and faults diagnosis. A new diagnosis process, combined with data mining and neural networks, are discussed as well as the functions and structure of agent which implements these algorithms. At last, some simulation results are shown to demonstrate the efficiency of the proposed system.

SENSOR-ASSISTED ADAPTIVE MOTOR CONTROL UNDER CONTINUOUSLY VARYING CONTEXT

Heiko Hoffmann, Georgios Petkos, Sebastian Bitzer and Sethu Vijayakumar

Adaptive motor control under continuously varying context, like the inertia parameters of a manipulated object, is an active research area that lacks a satisfactory solution. Here, we present and compare three novel strategies for learning control under varying context and show how adding tactile sensors may ease this task. The first strategy uses only dynamics information to infer the unknown inertia parameters. It is based on a probabilistic generative model of the control torques, which are linear in the inertia parameters. We demonstrate this inference in the special case of a single continuous context variable -- the mass of the manipulated object. In the second strategy, instead of torques, we use tactile forces to infer the mass in similar way. Finally, the third strategy omits this inference -- which may be infeasible if the latent space is multi-dimensional -- and directly maps the state, state transitions, and tactile forces onto the control torques. The additional tactile input implicitly contains all control-torque relevant properties of the manipulated object. In simulation, we demonstrate that this direct mapping can provide accurate control torques under multiple varying context variables.

SETPOINT ASSIGNMENT RULES BASED ON TRANSFER TIME DELAYS FOR WATER-ASSET MANAGEMENT OF NETWORKED OPEN-CHANNEL SYSTEMS

Eric Duviella, Pascale Chiron and Philippe Charbonnaud

The paper presents a new strategy based on a supervision and hybrid control accommodation to improve the water-asset management of networked open-channel systems. This strategy requires a modelling method of the network based on a weighted digraph of instrumented points, and the definition of resource allocation and setpoint assignment rules. Two setpoint assignment rules are designed and evaluated in the case of an open-channel system composed of one difluent and one confluent showing their effectiveness.

DISTRIBUTED CONTROL ARCHITECTURE FOR AUTOMATED NANOHANDLING

Christian Stolle

New distributed control architecture for micro- and nanohandling cells is presented. As a modular system it is designed to handle micro- and nanorobotic automation tasks at semi- up to full automation level. The architecture includes different visual sensors as there are scanning electron microscopes (SEM) and CCD cameras for position tracking as well as non-optical force, temperature, etc. sensors for environmental control. It allows usage of multiple nanorobots in parallel for combined autonomous fabrication tasks. The system provides a unified framework for mobile platforms and linear actors.

MODELING WITH CURRENT DYNAMICS AND VIBRATION CONTROL OF TWO PHASE HYBRID STEPPING MOTOR IN INTERMITTENT DRIVE

Ryota Mori, Yoshiyuki Noda, Takanori Miyoshi, Kazuhiko Terashima, Masayuki Nishida and Naohiko Suganuma

This paper presents modeling of stepping motor and control design of input pulse timing for the suppression control of vibration. The stepping motor has the transient response of electric current for the pulse input. Therefore, the motor model considering the transient response of the current is built. The validity of the proposed model is verified by comparing the model considering the transient response of the current with the one without its consideration. Design of the pulse input timing in the method of the four pulse drive is realized to achieve the desired angle without vibration and overshoot using an optimization method. Finally, the effectiveness of the proposed method is demonstrated by comparing simulation results with experiments.

PIECEWISE CONSTANT REINFORCEMENT LEARNING FOR ROBOTIC APPLICATIONS

Andrea Bonarini, Alessandro Lazaric and Marcello Restelli

Writing good behaviors for mobile robots is a hard task that requires a lot of hand tuning and often fails to consider all the possible configurations that a robot may face. By using reinforcement learning techniques a robot can improve its performance through a direct interaction with the surrounding environment and adapt its behavior in response to some non-stationary events, thus achieving a higher degree of autonomy with respect to pre-programmed robots. In this paper, we propose a novel reinforcement learning approach that addresses the main issues of learning in real-world robotic applications: experience is expensive, explorative actions are risky, control policy must be robust, state space is continuous. Preliminary results performed on a real robot suggest that on-line reinforcement learning, matching some specific solutions, can be effective also in real-world physical environments.

NONLINEAR PROGRAMMING IN APPROXIMATE DYNAMIC PROGRAMMING - Bang-bang Solutions, Stock-management and Unsmooth Penalties

Olivier Teytaud and Sylvain Gelly

Many stochastic dynamic programming tasks in continuous action-spaces are tackled through discretization. We here avoid discretization; then, approximate d ynamic programming (ADP) involves (i) many learning tasks, performed here by Support Vector Machines, for Bellman-function-regression (ii) many non-linear-o ptimization tasks for action-selection, for which we compare many algorithms. We include discretizations of the domain as particular non-linear-programming- tools in our experiments, so that by the way we compare optimization approaches and discretization methods. We conclude that robustness is strongly required in the non-linear-optimizations in ADP, and experimental results show that (i) discretization is sometimes inefficient, but some specific discretization is very efficient for "bang-bang" problems (ii) simple evolutionary tools outperform quasi-random in a stable manner (iii) gradient-based techniques are much less stable (iv) for most high-dimensional "less unsmooth" problems Covariance-Matrix-Adaptation is first ranked.

ACTIVE LEARNING IN REGRESSION, WITH APPLICATION TO STOCHASTIC DYNAMIC PROGRAMMING

Olivier Teytaud, Sylvain Gelly and Jérémie Mary

We study active learning as a derandomized form of sampling. We show that full derandomization is not suitable in a robust framework, propose partially derandomized samplings, and develop new active learning methods (i) in which expert knowledge is easy to integrate (ii) with a parameter for the exploration/exploitation dilemma (iii) less randomized than the full-random sampling (yet also not deterministic). Experiments are performed in the case of regression for value-function learning on a continuous domain. Our main results are (i) efficient partially derandomized point sets (ii) moderate-derandomization theorems (iii) experimental evidence of the importance of the frontier (iv) a new regression-specific user-friendly sampling tool less-robust than blind samplers but that sometimes works very efficiently in large dimensions. All experiments can be reproduced by downloading the source code and running the provided command line.

DC MOTOR FAULT DIAGNOSIS BY MEANS OF ARTIFICIAL NEURAL NETWORKS

Krzysztof Patan, Józef Korbicz and Gracjan Głowacki

The paper deals with a model-based fault diagnosis for a DC motor realized using artificial neural networks. Modelling of the considered process was carried out by using a neural network composed of dynamic neuron models. Decision making about possible faults was performed using statistical analysis of a residual. A neural network was applied to density shaping of a residual, and after that, assuming a significance level, a threshold was calculated. Moreover, to isolate faults a neural classifier was developed. The proposed approach was tested in DC motor laboratory systems at the nominal operations condition as well as in the case of faults.

HEURISTIC ALGORITHMS FOR SCHEDULING IN A MULTIPROCESSOR TWO-STAGE FLOWSHOP WITH 0-1 RESOURCE REQUIREMENTS

Ewa Figielska

This paper deals with the problem of preemptive scheduling in a two-stage flowshop with parallel unrelated machines at the first stage and a single machine at the second stage. At the first stage, jobs use some additional resources which are available in limited quantities at any time. The resource requirements are of 0-1 type. The objective is the minimization of makespan. The problem is NP-hard. Heuristic algorithms are proposed which, while solving to optimality the resource constrained scheduling problem at the first stage of the flowshop, select for simultaneous processing jobs according to rules promising a good (short) schedule in the flowshop. Several rules of job selection are considered. The performance of the proposed heuristic algorithms is analyzed by comparing their results with the lower bound on the optimal makespan. The results of computational experiments show that these heuristics are able to produce near-optimal solutions in short computational time.

AN INTELLIGENT MARSHALING PLAN BASED ON MULTI-POSITIONAL DESIRED LAYOUT IN CONTAINER YARD TERMINALS

Yoichi Hirashima

This paper proposes a new scheduling method for a marshaling in the container yard terminal. The proposed method is derived based on Q-Learning algorithm considering the desired position of containers that are to be loaded into a ship. In the method, 3 processes can be optimized simultaneously: rearrangement order of containers, layout of containers assuring explicit transfer of container to the desired position, and removal plan for preparing the rearrange operation. Moreover, the proposed method generates several desired positions for each container, so that the learning performance of the method can be improved as compared to the conventional methods. In general, at container yard terminals, containers are stacked in the arrival order. Containers have to be loaded into the ship in a certain order, since each container has its own shipping destination and it cannot be rearranged after loading. Therefore, containers have to be rearranged from the initial arrangement into the desired arrangement before shipping. In the problem, the number of container-arrangements increases by the exponential rate with increase of total count of containers, and the rearrangement process occupies large part of total run time of material handling operation at the terminal. For this problem, conventional methods require enormous time and cost to derive an admissible result. In order to show effectiveness of the proposed method, computer simulations for several examples are conducted.

SCHEDULING OF MULTI-PRODUCT BATCH PLANTS USING REACHABILITY ANALYSIS OF TIMED AUTOMATA MODELS

Subanatarajan Subbiah, Sebastian Panek, Sebastian Engell and Olaf Stursberg

Standard scheduling approaches in process industries are often based on algebraic problem formulations solved as MI(N)LP optimization problems to derive production schedules. To handle such problems techniques based on timed automata have emerged recently. This contribution reports on a successful application of a new modeling scheme to formulate scheduling problems in process industries as timed automata (TA) models and describes the solution technique to obtain schedules using symbolic reachability analysis. First, the jobs, resources and additional constraints are modeled as sets of synchronized timed automata. Then, the individual automata are composed by parallel composition to form a global automaton which has an initial location where no jobs have been started and at least one target location where all jobs have been finished. A cost optimal symbolic reachability analysis is performed on the composed automaton to derive schedules. The main advantage of this approach over other MILP techniques is the intuitive graphical and modular modeling and the ability to compute better solutions within reasonable computation time. This is illustrated by a case study.

AUTOMATIC ESTIMATION OF PARAMETERS FOR THE HIERARCHICAL REDUCTION OF RULES OF COMPLEX FUZZY CONTROLLERS

Yulia Ledeneva, Carlos A. Reyes-García and Alejandro Díaz-Méndez

Fuzzy control is an imitation of the fuzzy control laws that human use, which are expressed in the form of rules. The application of fuzzy control systems are of great importance in industry, navigation of space vehicles, flight control, missile speed control, etc. Frequently, such systems have many variables to control and are known as complex systems. For such systems, the fuzzy rule bases exponentially explode. The hierarchical method solves this problem by considerably reducing the number of rules. However, the performance of the resulting reduced system depends on the choice of some parameters which currently are found based on the experience and knowledge of a skilled system designer. In this work, we propose a method that uses a genetic algorithm to automatically estimate the corresponding parameters for the hierarchical reduction of the rule base. The implementation process, the simulation experiments and some results are presented.

ENHANCING KAPPA NUMBER CONTROL IN DOWNFLOW LO-SOLIDSTM DIGESTER USING DIAGNOSIS AND MODELLING

Timo Ahvenlampi and Rami Rantanen

In this study, Kappa number prediction and diagnosis in continuous Downflow Lo-Solids$^{TM}$ cooking application is investigated. The Kappa number is one of the quality measures in the pulp cooking process and usually the only on-line measurement. It is a measure of the residual lignin content in the pulp. The Kappa number is mainly controlled by the cooking temperature. In this study, Kappa number control (temperature control) is carried out using Gustafson's Kappa number model for the prediction of the blowline Kappa number. New temperature set point is solved iteratively based on the difference between the predicted and target blow-line kappa numbers. The input variables are monitored using self-organizing map (SOM). The data is collected from industrial continuous Downflow Lo-Solids{TM} cooking digester. Good results were achieved using the proposed approach.

GLOBAL ASYMPTOTIC VELOCITY OBSERVATION OF NONLINEAR SYSTEMS - Application to a Frictional Industrial Emulator

R. Guerra, C. Iurian, L. Acho, F. Ikhouane and J. Rodellar

In mechanical systems with friction, development of velocity observers deserves a special emphasis because, as evidenced in numerical and experimental tests when a state-of-the-art observer is armed, friction can induce high-frequency oscillations in the estimated velocity. In this short paper, two new velocity-observation algorithms are designed, based on this previously reported observer, which eliminate the high-frequency oscillations noted in the original one. Numerical and experimental performance comparisons are carried through in a mechanical PID control system where the estimated velocity is incorporated into the feedback loop.

A MULTI AGENT CONTROLLER FOR A MOBILE ARM MANIPULATOR

Sébatien Delarue, Philippe Hoppenot and Etienne Colle

In the assistive robotics domain, and especially for disable people, the use of mobile arm manipulator can be of a great help in the everyday life tasks. First, these systems must be reliable and fault tolerant. Secondly they must facilitate man machine co-operation. This article exposes a method based on multi agent system. This kind of distributed architecture makes possible to be fault-tolerant without any specific fault management, and thus to improve reliability. It is also possible to add specific constraints, for example human like behaviors in order to facilitate the use of the system by the operator. Moreover, this method is easy to implement

A PARAMETERIZED GENETIC ALGORITHM IP CORE DESIGN AND IMPLEMENTATION

K. M. Deliparaschos, G. C. Doyamis and S. G. Tzafestas

Genetic Algorithm (GA) is a directed random search technique working on a population of solutions and based on natural selection. However, its convergence to the optimum may be very slow for complex optimization problems, especially when the GA is software implemented, making it difficult to be used in real time applications. In this paper a parameterized GA IP is designed and implemented on hardware, achieving impressive time–speedups when compared to its software version. The parameterization stands for the number of population individuals and their bit resolution, the bit resolution of each individual’s fitness, the number of elite genes in each generation, the crossover and mutation methods, the maximum number of generations, the mutation probability and its bit resolution. The proposed architecture is implemented in a field programmable gate array chip (FPGA) with the use of a very high-speed integrated circuits hardware description language (VHDL) and advanced synthesis and place and route tools. The GA discussed in this work achieves a frequency rate of 92 MHz and is evaluated using the Traveling Salesman Problem as well as several benchmarking functions.

TRACKING A WHEELCHAIR WITH A MOBILE PLATFORM

B.Allart, B. Marhic, L. Delahoche, A. Clérentin and O. Rémy-Néris

This article deals with a target tracking application for the disabled. The objective of this work is to track a wheelchair with a mobile platform and an embedded grasping arm (MANUS). We propose an approach based on an association of two Kalman filtering levels. The first level permits an estimation of the wheelchair configuration. The second is used to compute the mobile platform configuration in connection with its environment. The association of the two filtering process allows a robust tracking between two objects in movement.

ON TUNING THE DESIGN OF AN EVOLUTIONARY ALGORITHM FOR MACHINING OPTIMIZATION PROBLEMS

Jean-Louis Vigouroux, Sebti Foufou1, Laurent Deshayes, James J. Filliben, Lawrence A. Welsch and M. Alkan Donmez

In this paper, a methodology for tuning the design of an evolutionary algorithm (EA) is presented. An EA for solving machining optimization problems having highly non-linear constraints and uncertainties is studied. A conventional turning optimization problem, solved previously with classic optimization algorithms, serves as a basis for the investigation of the EA. The parameters of the problem now can be modified in a certain range, and statistical engineering methods are used to find a unique set of algorithm parameters giving robust results.

RSRT: RAPIDLY EXPLORING SORTED RANDOM TREE - Online Adapting RRT to Reduce Computational Solving Time while Motion Planning in Wide Configuration Spaces

Nicolas Jouandeau

We present a new algorithm, named RSRT, for Rapidly-exploring Random Trees(RRT) based on inherent relations analysis between RRT components. RRT algorithms are designed to consider interactions between these inherent components. We explain properties of known variations and we present some future once which are required to deal with dynamic strategies. We present experimental results for a wide set of path planning problems involving a free flying object in a static environment. The results show that our RSRT algorithm is faster than existing ones. This results can also stand as a starting point of a motion planning benchmark instances which would make easier further comparative studies of path planning algorithms.

THE VERIFICATION OF TEMPORAL KNOWLEDGE BASED SYSTEMS - A Case-study on Power-systems

Jorge Santos, Zita Vale, Carlos Ramos and Carlos Serôdio

The verification and validation (V\&V) process states if the software requirements specifications have been correctly and completely fulfilled. The methodologies proposed in software engineering showed to be inadequate for knowledge based systems (KBS) validation and verification, since KBS present some particular characteristics. Designing KBS for dynamic environments requires the consideration of temporal knowledge reasoning and representation (TRR) issues. Although humans present a natural ability to deal with knowledge about time and events, the codification and use of such knowledge in information systems still pose many problems. Hence, the development of applications strongly based on temporal reasoning remains an hard and complex task. Furthermore, albeit the last significant developments in TRR area, there is still a considerable gap for its successful use in practical applications. VERITAS is an automatic tool developed for KBS verification which is able to detect a large number of knowledge anomalies. It addresses many relevant aspects considered in real applications, like the usage of rule triggering selection mechanisms and temporal reasoning. This paper presents a solution, based in the combination of formal methods and heuristics, addressing some open issues on verification of KBS applied in critical domains.

A COMPARISON OF HUMAN AND MARKET-BASED ROBOT TASK PLANNERS

Guido Zarrella, Robert Gaimari and Bradley Goodman

Urban search and rescue, reconnaissance, manufacturing, and team sports are all problem domains requiring multiple agents that are able to collaborate intelligently to achieve a team goal. In these domains task planning and assignment can be challenging to robots and humans alike. In this paper we introduce a market-based distributed task planning algorithm that has been adapted for heterogeneous, tightly coordinated robots in domains with time deadlines. We also report the results of our experiments comparing the robots' decisions with the decisions produced by ten teams of humans performing an identical search and rescue task. The outcome provides insight into the types of problems for which information technology can add value by providing decision support for human problem solvers.

HOLONIC PRODUCTION PROCESS: A MODEL OF COMPLEX, PRECISE, AND GLOBAL SYSTEMS

Edgar Chacon, Isabel Besembel, Dulce Rivero and Juan Cardillo

The abstract should summarize the contents of the paper and should contain at least 70 and at most 200 Nowadays, it is needed a complete description of the production process in order to plan, program, control, and supervise the production process itself. The complexity to obtain this description is due to the integration of two contradictory points of views. First, the precision implicated in the construction of total and complete models, and on the other hand, the need of having a global vision associated with the different views of the process. These views normally show three important aspects: the structural organization of the model, the dynamism between the main components, and the distinct temporal scales and levels, where are taken the main decisions. The holonic approach (Erikson,2004) has been used to manage this complexity, in order to have an abstraction that permit the integration of the mentioned points of views.

CHANGE-POINT DETECTION WITH SUPERVISED LEARNING AND FEATURE SELECTION

Victor Eruhimov, Vladimir Martyanov, Eugene Tuv and George C. Runger

Data streams with high dimensions are more and more common as data sets become wider. Time segments of stable system performance are often interrupted with change events. The change-point problem is to detect such changes and identify attributes that contribute to the change. Existing methods focus on detecting a single (or few) change-point in a univariate (or low-dimensional) process. We consider the important highdimensional multivariate case with multiple change-points and without an assumed distribution. The problem is transformed to a supervised learning problem with time as the output response and the process variables as inputs. This opens the problem to a wide set of supervised learning tools. Feature selection methods are used to identify the subset of variables that change. An illustrative example illustrates the method in an important type of application.

MULTICRITERIAL DECISION-MAKING IN ROBOT SOCCER STRATEGIES

Petr Tucnık, Jan Kozany and Vilém Srovnal

The principle of multicriterial decision-making is used for the purpose of autonomous control of both individual agent and the multiagent team as a whole. This approach to the realization of control mechanism is non-standard and experimental and the robot soccer game was chosen as a testing ground for this control method. It provides an area for further study and research and some of the details of its design will be presented in this paper.

MINIMIZING THE ARM MOVEMENTS OF A MULTI-HEAD GANTRY MACHINE

Timo Knuutila, Sami Py¨otti¨al¨a and Olli S. Nevalainen

In printed circuit board (PCB) manufacturing multi-head gantry machines are becoming increasingly more popular in surface mount technology (SMT), because they combine high speed with moderate price. This kind of machine picks up several components from the feeder and places them on the PCB. The process is repeated until all component placements are done. In this article, a subproblem of the machine control is studied. Here, the placement order of the components, the nozzles in the placement arm and the component locations in the feeder are fixed. The goal is to find an optimal pick-up sequence when minimizing the total length of the arm movements. An algorithm that searches the optimal pick-up sequence is proposed and tested widely. Tests show that the method can be applied to problems of practical size.

A GROWING FUNCTIONAL MODULE DESIGNED TO TRIGGER CAUSAL INFERENCE

Jérôme Leboeuf Pasquier

“Growing Functional Modules” constitutes a prospective paradigm founded on the epigenetic approach whose proposal consists in designing a distributed architecture, based on interconnected modules, that allows the automatic generation of an autonomous and adaptive controller (artificial brain). The present paper introduces a new module designed to trigger causal inference; its functionality is discussed and its behavior is illustrated applying the module to solve the problem of a dynamic maze.

A MULTI CRITERIA EVALUATION OVER A FINITE SCALE FOR MAINTENANCE ACTIVITIES OF A MOTORWAY OPERATOR

Céline Sanchez, Jacky Montmain, Marc Vinches and Brigitte Mahieu

The Escota Company aims at the formalization and improvement of the decisional process for preventive maintenance in a multi criteria (MC) environment. According to available pieces of knowledge on the infrastructure condition, operations are to be evaluated with regards to (w.r.t.) technical but also to conformity, security and financial criteria. This MC evaluation is modelled as the aggregation of partial scores attributed to an operation w.r.t. a given set of n criteria. The scores are expressed over a finite scale which can cause some troubles when no attention is paid to the aggregation procedure. This paper deals with the consistency of the evaluation process, where scores are expressed as labels by Escota’s experts, whereas the aggregation model is supposed to deal with numerical values and cardinal scales. We try to analyse this curious but common apparent paradox in MC evaluation when engineering contexts are concerned. A robustness study of the evaluation process concludes this paper.

COGNITIVE APPROACH TO PROBLEM SOLVING OF SOCIAL AND ECONOMIC OBJECT DEVELOPMENT

Z. Avdeeva, S. Kovriga and D. Makarenko

The basic technique of problem-solving is structurization of knowledge about object and its environment and construction of a cognitive model. The technique includes monitoring of dynamics of factors of the model (their tendencies), analysis of the model structure with the use of SWOT-approach, and modeling that permits to determine and solve semi-structured problems. The technique allows supporting of a vital control task that consists in goal setting of socio-economic object development, as far as solution of discovered problems turns into the system development control task. The application of technique is useful when designing a strategy of development of social and economic objects.

FEASIBILITY OF SUBSPACE IDENTIFICATION FOR BIPEDS - An Innovative Approach for Kino-Dynamic Systems

Muhammad Saad Saleem and Ibrahim A. Sultan

Different approaches have been overviewed which have been used in stability of biped robots. Current implementations either mimic human behavior or use heuristic control. This paper suggests the use of supervisory crisp control in operational space configuration for better control and understanding of kino-dynamic systems and biped robots.

IDENTIFICATION OF MODELS OF EXTERNAL LOADS

Yuri Menshikov

In the given work the problem of construction (synthesis) of mathematical model of unknown or little-known external load (EL) on open dynamic system is considered. Such synthesis is carried out by special processing of the experimentally measured response of dynamic system on researched real external load and known external loads (method of identification). This problem is considered in two statements: the synthesis of EL for certain model and the synthesis of EL