DOI: 10.3897/jucs.116480

Authors: Besma Hezili, Hichem Talbi

Abstract: We present a Collaborative Auto-Diversified Optimization Scheme (CADOS) for solving continuous and combinatorial optimization problems. CADOS aims to explore the synergy of various optimization algorithms and enhance their effectiveness and efficiency, particularly in higher-dimensional problems. It incorporates an enhanced version of the previously proposed approach Auto-Diversified Ameliorated MultiPopulation-based Ensemble Differential Evolution (AD-AMPEDE), Covariance Matrix Adaptation Evolutionary Strategy (CMA-ES), and a local search (LS) algorithm. AD-AMPEDE has demonstrated good performance in solving continuous optimization problems. However, its competitiveness wanes in higher dimensions. CADOS improves AD-AMPEDE’s detection/re-diversification processes and parameters adaptation, making it effective for higher-dimensional problem classes. To explore nearby regions during stagnation, a trust-region local search is employed. For re-diversification, CADOS utilizes both CMA-ES, known for its efficiency in complex fitness landscapes, and the Auto-Enhanced Population Diversity (AEPD) technique. We tested CADOS on the COmparing Continuous Optimizers (COCO) platform and the results demonstrated excellent performance of CADOS. In addition, to show the proposed scheme’s efficacy in tackling real-world issues, we employed it to optimize the design of water distribution networks (WDS). The results we obtained underscore the remarkable competitiveness of our strategy when compared to widely recognized existing algorithms.

HTML

PDF

DOI: 10.3897/jucs.70341

Authors: Lamia Berkani, Lylia Betit, Louiza Belarif

Abstract: Clustering-based approaches have been demonstrated to be efficient and scalable to large-scale data sets. However, clustering-based recommender systems suffer from relatively low accuracy and coverage. To address these issues, we propose in this article an optimized multiview clustering approach for the recommendation of items in social networks. First, the selection of the initial medoids is optimized using the Bees Swarm optimization algorithm (BSO) in order to generate better partitions (i.e. refining the quality of medoids according to the objective function). Then, the multiview clustering (MV) is applied, where users are iteratively clustered from the views of both rating patterns and social information (i.e. friendships and trust). Finally, a framework is proposed for testing the different alternatives, namely: (1) the standard recommendation algorithms; (2) the clustering-based and the optimized clustering-based recommendation algorithms using BSO; and (3) the MV and the optimized MV (BSO-MV) algorithms. Experimental results conducted on two real-world datasets demonstrate the effectiveness of the proposed BSO-MV algorithm in terms of improving accuracy, as it outperforms the existing related approaches and baselines.

HTML

XML

PDF

DOI: 10.3897/jucs.2020.040

Authors: Łukasz Apiecionek, Jacek Czerniak, Dawid Ewald, Mateusz Biedziak

Abstract: There is currently an era of Internet of Things in the computer systems, which consists in connecting all possible devices to the Internet in order to provide them with new functionalities and thus { to improve the user's life standard. One of such solutions could be Smart Home. The possibility of monitoring inner environment is required for such solutions. Such monitoring provides potential for e.g. better heating control. The authors of this paper propose some heating control method with Fuzzy Logic. The proposed method was tested in a special climate chamber. The authors provided conclusions at the end of the paper.

HTML

XML

PDF

DOI: 10.3897/jucs.2020.016

Authors: Murilo Falleiros Lemos Schmitt, Mauro Mulati, Ademir Constantino, Fábio Hernandes, Tony Hild

Abstract: This paper proposes an algorithm for the set covering problem based on the metaheuristic Ant Colony Optimization (ACO) called Ant-Set, which uses a lineoriented approach and a novelty pheromone manipulation based on the connections between components of the construction graph, while also applying a local search. The algorithm is compared with other ACO-based approaches. The results obtained show the effectiveness of the algorithm and the impact of the pheromone manipulation.

HTML

XML

PDF

DOI: 10.3217/jucs-025-10-1221

Authors: K. Shankar, Mohamed Elhoseny

Abstract: In wireless communication, Mobile Ad Hoc Network (MANET) consists of a number of mobile nodes which are communicated with each other without any base station. One of the security attacks in MANETs is Packet forwarding misbehaviour attack; this makes MANETs weak by showing message loss behavior. For securing message transmission in MANET, the work proposes Energy Efficient Clustering Protocol (EECP) with Radial Basis Function (RBF) based CH is elected for formed Clusters. Moreover, here some Network measures are considered to detect the malicious nodes and CH model that is speed, mobility, trust and so on. The trust value of the node is computed from the neighbor node which helps in further location to find a malicious node in the network to avail message drop and energy consumption (EC). After detecting malicious nodes, Multi secure Protocols that is Secure Efficient Distance Vector Routing (SEDV) and Secure Link State Routing Protocol (SLSP) with encryption technique used for message security. If the" HELLO" message sending by the sender, its encrypted and decrypted triples in receiver end to get the plain message, this technique is Triple Data Encryption Standard (TDES). Finally, the implementation results are evaluated to analyze the message security level of the proposed system in MANET in terms, of Packet to Delivery Ratio (PDR, Network Life Time (NLT) and some other important Measures.

HTML

XML

PDF

DOI: 10.3217/jucs-024-09-1271

Authors: Yuyong Jeon, Ilkyeun Ra, Youngjin Park, Sangmin Lee

Abstract: In this paper, a fast and effective method of parameter optimization for noise estimation is proposed for various types of noise. The proposed method is based on gradient descent, which is one of the optimization methods used in machine learning. The learning rate of gradient descent was set to a negative value for optimizing parameters for a speech quality improvement problem. The speech quality was evaluated using a suite of measures. After parameter optimization by gradient descent, the values were re-checked using a wider range to prevent convergence to a local minimum. To optimize the problem's five parameters, the overall number of operations using the proposed method was 99.99958% smaller than that using the conventional method. The extracted optimal values increased the speech quality by 1.1307%, 3.097%, 3.742%, and 3.861% on average for signal-to-noise ratios of 0, 5, 10, and 15 dB, respectively.

HTML

XML

PDF

DOI: 10.3217/jucs-024-06-0682

Authors: Asit Das, Soumen Pati, Hsien-Hung Huang, Chi-Ken Chen

Abstract: Microarray dataset contains huge number of genes, many of which are irrelevant regarding cancer classification and as a result classification accuracy is reduced. Therefore, the dataset should be pre-processed to filter out these redundant genes. In this paper, initially a Pareto optimality based Multi-objective Genetic Algorithm has been proposed where non-linear cellular automata is employed to overcome the demerits of random initialization to generate initial population in high dimensional space. The fitness functions are defined based on both attribute dependency and boundary region exploration of rough set theory and Log-Likelihood ratio to select the informative genes. The chromosomes are hybridized by applying multi-point crossover; whereas proximity mutation builds on Flip-bit mutation with a little modification to produce fittest offspring. Finally, the gene subset with strong biological significance in cancer treatment is obtained from the Pareto dominant solutions. Performances are investigated on publicly available microarray cancer datasets and compared with the state-of-the-art methods to demonstrate the effectiveness of the proposed method.

HTML

XML

PDF

DOI: 10.3217/jucs-022-10-1298

Authors: Virginia Yannibelli, Marcelo Armentano, Franco Berdun, Anala Amandi

Abstract: Collaborative learning team building is a fundamental, difficult and time-consuming task in educational environments. In this paper, we address a collaborative learning team building problem that considers two valuable grouping criteria usually considered by teachers. One of these criteria considers the understanding levels of the students with respect of the topics of a given course, and is based on building well-balanced teams in terms of the understanding levels of their members. The other criterion considers the interest levels of the students with respect of the topics of a given course, and is based on building well-balanced teams in terms of the interest levels of their members. The problem addressed has been recognised as an NP-Hard optimization problem. To solve the problem, we propose a steady-state evolutionary algorithm. This algorithm aims to organize the students taking a given course into teams in such a way that the two grouping criteria of the problem are optimized. The performance of the algorithm is evaluated on nine problem instances with different levels of complexity, and is compared with that of the only algorithm previously proposed for solving the addressed problem. The obtained results show that the steady-state evolutionary algorithm significantly outperforms the previous algorithm.

HTML

XML

PDF

DOI: 10.3217/jucs-022-06-0874

Authors: Piotr Jędrzejowicz, Aleksander Skakovski

Abstract: Differential evolution (DE) is a stochastic global optimization method, that has been under continuous development during the past two decades. It has been recognized that preserving the diversification of population can significantly improve the performance of DE. Although, several results and approaches to population diversification have been proposed, it seems that this issue still has a potential for development. In this paper we have studied experimentally the possibility of increasing the performance of DE. Our investigation aims at identifying how the performance of DE depends on such factors as population diversity, size and number of fitness function evaluations carried out by DE to yield a solution. In our experiments we diversified the population in an intensive manner using the proposed decloning procedure carried out in cycles, and also through increasing the population size. The choice of how to preserve the diversification may depend on restrictions imposed on the population size, response time, and the quality of solutions that should be met by a specific implementation of the algorithm. The obtained results allowed us to propose a performance improvement policy that might noteworthy improve both the efficacy and response time of the algorithm. The discrete-continuous scheduling with continuous resource discretisation was used as the test problem.

HTML

XML

PDF

DOI: 10.3217/jucs-021-04-0503

Authors: Robert Bucki, Bronislav Chramcov, Petr Suchánek

Abstract: The paper highlights the problem of minimizing economic costs of making orders in the automated manufacturing system which consists of work centres arranged in a series. Each of them is equipped with tools which carry out defined manufacturing operations. Tools are replaced with new ones only when no manufacturing operation can be performed any more in order to minimize the residual pass. The equations of state of the production line are presented and heuristic control strategies are discussed in detail. The criterion is to minimize the number of replacement procedures which results in maximizing the use of tools in work centres. To prove the correctness of the presented approach the paper is supported with an extended simulation study based on implementing available combinations of either manufacturing or replacement strategies taking into account various configurations which come into being in the real manufacturing environment. The simulation results form the basis for the detailed analysis to meet the requirements of the applicable decision-making procedures.

HTML

XML

PDF

DOI: 10.3217/jucs-020-13-1855

Authors: Ming Yang, Bo Liu, Wei Wang, Junzhou Luo, Xiaojun Shen

Abstract: By exploiting multi-radio multi-channel technology, wireless mesh networks can effectively provide wireless broadband access to the Internet for mobile users. Due to the limited number of orthogonal channels, overlapping channel assignment is one of the main factors that greatly affect the network capacity. However, current results in this area are not so satisfying. In this paper, we first propose a model for measuring achieved network capacity in MR-WMNs. Then we prove that finding an optimal overlapping channel assignment in a given MR-WMN with odd number of channels, is equivalent to finding an optimal assignment by only using its orthogonal channels. This theory allows us to use fewer channels to solve complicated channel assignment problems. Third, we prove that in 802.11b/g MR-WMN the simplified optimization problem is a Max-3-Cut problem. Although this problem is NP-hard, it has an efficient approximation algorithm that achieves approximation ratio of 1.19616 probabilistically by using the algorithm for Max-Cut whose approximation ratio is 1.1383 probabilistically. Based on the algorithm for Max-Cut, this paper proposes Max-Cut based channel assignment (MCCA) which uses a heuristic method to adjust the result produced by the Max-Cut algorithm to achieve an even better result. Finally, we perform extensive simulations to compare the MCCA with a state-of-the-art Tabu-Search based algorithm. The results show that the Max-Cut based overlapping channel assignment algorithm effectively and efficiently improves on the network capacity compared with existing algorithms.

HTML

XML

PDF

DOI: 10.3217/jucs-018-14-1906

Authors: Antonio Bajuelos, Santiago Canales, Gregorio Hernández, Mafalda Martins

Abstract: In this paper we focus on approximate solutions to solve a new class of Art Gallery Problems inspired by wireless localization. Instead of the usual guards we consider wireless devices whose signal can cross a certain number, k, of walls. These devices are called k-transmitters. We propose an algorithm for constructing the visibility region of a k-transmitter located on a point of a simple polygon. Then we apply a hybrid metaheuristic strategy to tackle the problem of minimizing the number of k-transmitters, located at vertices, that cover a given simple polygon, and compare its performance with two pure metaheuristics. We conclude that the approximate solutions obtained with the hybrid strategy, for 2-transmitters and 4-transmitters, on simple polygons, monotone polygons, orthogonal polygons and monotone orthogonal polygons, are better than the solutions obtained with the pure strategies.

HTML

XML

PDF

DOI: 10.3217/jucs-018-13-1853

Authors: Musrrat Ali, Millie Pant, Atulya Nagar, Chang Ahn

Abstract: Insertion of a local search technique is often considered an effective mechanism to increase the efficiency of a global optimization algorithm. In this paper we propose and analyze the effect of two local searches namely; Trigonometric Local Search (TLS) and Interpolated Local Search (ILS) on the working of basic Differential Evolution (DE). The corresponding algorithms are named as DETLS and DEILS. The performances of proposed algorithms are investigated and compared with basic DE, modified versions of DE and some other evolutionary algorithms. It is found that the proposed schemes improve the performance of DE in terms of quality of solution without compromising with the convergence rate.

HTML

XML

PDF

DOI: 10.3217/jucs-018-13-1842

Authors: Kusum Deep, Jagdish Bansal

Abstract: Particle Swarm Optimization (PSO) is a swarm intelligence optimization method inspired from birds' flocking or fish schooling. Many improved versions of PSO are reported in literature, including some by the authors. Original as well as improved versions of PSO have proven their applicability to various fields like science, engineering and industries. Economic dispatch (ED) problem is one of the fundamental issues in power system operations. This problem turns out to be a non linear continuous optimization problem. In this paper, economic dispatch problem is solved using original PSO and two of its improved variants, namely, Laplace Crossover PSO (LXPSO) and Quadratic Approximation PSO (qPSO), in order to find better results than reported in the literature. Results are also compared with the earlier published results.

HTML

XML

PDF

DOI: 10.3217/jucs-017-14-2064

Authors: Junzhou Luo, Wenjia Wu, Ming Yang

Abstract: In a wireless mesh network (WMN), gateways act as the bridges between the mesh backbone and the Internet, and significantly affect the performance of the whole network. Hence, how to determine the optimal number and positions of gateways, i.e., gateway deployment, is one of the most important and challenging topics in practical and theoretical research on designing a WMN. Although several approaches have been proposed to address this problem, few of them take load balancing and interference minimization into account. In this paper, we study the Load-balancing and Interference-minimization Gateway Deployment Problem (LIGDP), which aims to achieve four objectives, i.e. minimizing deployment cost, minimizing MR-GW path length, balancing gateway load and minimizing link interference. We formulate it as a multi-objective integer linear program (ILP) issue first, and then propose an efficient gateway deployment approach, called LIGDP Heuristic. The approach joints two heuristic algorithms, i.e., MSC-based location algorithm (MLA) and load-aware and interference-aware association algorithm (LIAA), to determine gateway positions and construct GW-rooted trees. Simulation results not only show that the trade-off between deployment cost and network performance can be achieved by adjusting R-hop, GW throughput and MR throughput constraints, but also demonstrate that, compared with other existing approaches, LIGDP Heuristic performs better on MR-GW path, load balancing and interference minimization without deploying more gateways.

HTML

XML

PDF

DOI: 10.3217/jucs-017-14-2048

Authors: Chen-Fang Tsai, Weidong Li, Anne James

Abstract: This paper presents a new methodology for improving the efficiency and generality of Genetic Algorithms (GA). The methodology provides the novel function of adaptive parameter adjustment during each evolution generation of GA. The important characteristics of the methodology are mainly from the following two aspects: (1) superior performance members in GA are preserved and inferior performance members are deteriorated to enhance search efficiency towards optimal solutions; (2) adaptive crossover and mutation management is applied in GA based on the transformation functions to explore wider spaces so as to improve search effectiveness and algorithm robustness. The research was successfully applied for a luggage design chain to generate optimal solutions (minimized lifecycle cost). Experiments were conducted to compare the work with the prior art to demonstrate the characteristics and advantages of the research.

HTML

XML

PDF

DOI: 10.3217/jucs-016-14-1882

Authors: Pablo Rabanal, Ismael Rodriguez, Fernando Rubio

Abstract: River Formation Dynamics (RFD) is an evolutionary computation methodbased on copying how drops form rivers by eroding the ground and depositing sediments. Given a cost-evaluated graph, we apply RFD to find a way to connect a givenset of origins with a given destination in such a way that distances from origins to the destination are minimized (thus improving the quality of service) but costs to build theconnecting infrastructure are minimized (thus reducing investment expenses). After we prove the NP-completeness of this problem, we apply both RFD and an Ant ColonyOptimization (ACO) approach to heuristically solve it, and some experimental results are reported.

HTML

XML

PDF

DOI: 10.3217/jucs-016-06-0983

Authors: Maytham Safar, Nosayba El-Sayed, Khaled Mahdi, David Taniar

Abstract: Recent work on social networks has tackled the measurement and optimization of these networks’ robustness and resilience to both failures and attacks. Different metrics have been used to quantitatively measure the robustness of a social network. In this work, we design and apply a Genetic Algorithm that maximizes the cyclic entropy of a social network model, hence optimizing its robustness to failures. Our social network model is a scale-free network created using Barabási and Albert's generative model, since it has been demonstrated recently that many large complex networks display a scale-free structure. We compare the cycles distribution of the optimally robust network generated by our algorithm to that belonging to a fully connected network. Moreover, we optimize the robustness of a scale-free network based on the links-degree entropy, and compare the outcomes to that which is based on cycles-entropy. We show that both cyclic and degree entropy optimization are equivalent and provide the same final optimal distribution. Hence, cyclic entropy optimization is justified in the search for the optimal network distribution.

HTML

XML

PDF

DOI: 10.3217/jucs-015-13-2446

Authors: Dunwei Gong, Xin Yao, Jie Yuan

Abstract: Interactive genetic algorithms (IGAs) are effective methods to solve optimization problems with implicit or fuzzy indices. But human fatigue problem, resulting from evaluation on individuals and assignment of their fitness, is very important and hard to solve in IGAs. Aiming at solving the above problem, an interactive genetic algorithm with an individual fitness not assigned by human is proposed in this paper. Instead of assigning an individual fitness directly, we record time to choose an individual from a population as a satisfactory or unsatisfactory one according to sensitiveness to it, and its fitness is automatically calculated by a transformation from time space to fitness space. Then subsequent genetic operation is performed based on this fitness, and offspring is generated. We apply this algorithm to fashion design, and the experimental results validate its efficiency.

HTML

XML

PDF

DOI: 10.3217/jucs-015-04-0826

Authors: Vitoantonio Bevilacqua, Giuseppe Mastronardi, Filippo Menolascina, Paolo Pannarale, Giuseppe Romanazzi

Abstract: Inferring gene regulatory networks from data requires the development of algorithms devoted to structure extraction. When time-course data is available, gene interactions may be modeled by a Bayesian Network (BN). Given a structure, that models the conditional independence between genes, we can tune the parameters in a way that maximize the likelihood of the observed data. The structure that best fit the observed data reflects the real gene network's connections. Well known learning algorithms (greedy search and simulated annealing) devoted to BN structure learning have been used in literature. We enhanced the fundamental step of structure learning by means of a classical evolutionary algorithm, named GA (Genetic algorithm), to evolve a set of candidate BN structures and found the model that best fits data, without prior knowledge of such structure. In the context of genetic algorithms, we proposed various initialization and evolutionary strategies suitable for the task. We tested our choices using simulated data drawn from a gene simulator, which has been used in the literature for benchmarking [Yu et al. (2002)]. We assessed the inferred models against this reference, calculating the performance indicators used for network reconstruction. The performances of the different evolutionary algorithms have been compared against the traditional search algorithms used so far (greedy search and simulated annealing). Finally we individuated as best candidate an evolutionary approach enhanced by Crossover-Two Point and Selection Roulette Wheel for the learning of gene regulatory networks with BN. We show that this approach outperforms classical structure learning methods in elucidating the original model of the simulated dataset. Finally we tested the GA approach on a real dataset where it reach 62% of recovered connections (sensitivity) and 64% of direct connections (precision), outperforming the other algorithms.

HTML

XML

PDF

DOI: 10.3217/jucs-015-04-0722

Authors: Xuesong Wang, Minglin Hao, Yuhu Cheng, Ruhai Lei

Abstract: Differential evolution (DE) algorithm puts emphasis particularly on imitating the microscopic behavior of individuals, while estimation of distribution algorithm (EDA) tries to estimate the probabilistic distribution of the entire population. DE and EDA can be extended to multi-objective optimization problems by using a Pareto-based approach, called Pareto DE (PDE) and Pareto EDA (PEDA) respectively. In this study, we describe a novel combination of PDE and PEDA (PDE-PEDA) for multi-objective optimization problems by taking advantage of the global searching ability of PEDA and the local optimizing ability of PDE, which can, effectively, maintain the balance between exploration and exploitation. The basic idea is that the offspring population of PDE-PEDA is composed of two parts, one part of the trial solution generated originates from PDE and the other part is sampled in the search space from the constructed probabilistic distribution model of PEDA. A scaling factor Pr used to balance contributions of PDE and PEDA can be adjusted in an on-line manner using a simulated annealing method. At an early evolutionary stage, a larger Pr should be adopted to ensure PEDA is used more frequently, whereas at later stage, a smaller Pr should be adopted to ensure that offspring is generated more often using PDE. The hybrid algorithm is evaluated on a set of benchmark problems and the experimental results show that PDE-PEDA outperforms the NSGA-II and PDE algorithms.

HTML

XML

PDF

DOI: 10.3217/jucs-014-15-2566

Authors: Khadiza Tahera, Raafat Ibrahim, Paul Lochert

Abstract: T his paper proposes a variant of genetic algorithm - GADYM, Genetic Algorithm with Gender-Age structure, DYnamic parameter tuning and Mandatory self perfection scheme. The motivation of this algorithm is to increase the diversity throughout the search procedure and to ease the difficulties associated with the tuning of GA parameters and operators. To promote diversity , GADYM combines the concept of gender and age in individuals of a traditional Genetic Algorithm and implements the self perfection scheme through sharing. To ease the parameter tuning process, the proposed algorithm uses dynamic environment in which heterogeneous crossover and selection techniques are used and parameters are updated based on deterministic rules. Thus, GADYM uses a combination of genetic operators and variable parameter values whereas a traditional GA uses fixed values of those. The experim ental results of the proposed algorithm based on a mechanical design problem show promising result.

HTML

XML

PDF

DOI: 10.3217/jucs-014-15-2549

Authors: Guillermo Molina, Enrique Alba, El-Ghazali Talbi

Abstract: Wireless Sensor Networks (WSN) allow, thanks to the use of small wireless devices known as sensor nodes, the monitorization of wide and remote areas with precision and liveness unseen to the date without the intervention of a human operator. For many WSN applications it is fundamental to achieve full coverage of the terrain monitored, known as sensor field. The next major concerns are the energetic efficiency of the network, in order to increase its lifetime, and having the minimum possible number of sensor nodes, in order to reduce the network cost. The task of placing the sensor nodes while addressing these objectives is known as WSN layout problem. In this paper we address a WSN layout problem instance in which full coverage is treated as a constraint while the other two objectives are optimized using a multiobjective approach. We employ a set of multi-objective optimization algorithms for this problem where we define the energy efficiency and the number of nodes as the independent optimization objectives. Our results prove the efficiency of multi-objective metaheuristics to solve this kind of problem and encourage further research on more realistic instances and more constrained scenarios.

HTML

XML

PDF

DOI: 10.3217/jucs-014-05-0766

Authors: Sergios Soursos, Costas Courcoubetis, Richard Weber

Abstract: Nowadays, in the markets of broadband access services, traditional contracts are of "static" type. Customers buy the right to use a specific amount of resources for a specific period of time. On the other hand, modern services and applications render the demand for bandwidth highly variable and bursty. New types of contracts emerge ("dynamic contracts") which allow customers to dynamically adjust their bandwidth demand. In such an environment, we study the case of a price competition situation between two providers of static and dynamic contracts. We investigate the resulting reaction curves, search for the existence of an equilibrium point and examine if and how the market is segmented between the two providers. Our first model considers simple, constant provision costs. We then extend the model to include costs that depend on the multiplexing capabilities that the contracts offer to the providers, taking into consideration the size of the market. We base our analysis on the theory of effective bandwidths and investigate the new conditions that allow the provider of dynamic contracts to enter the market.

HTML

XML

PDF

DOI: 10.3217/jucs-012-09-1385

Authors: József Tick, Zoltán Kovacs, Ferenc Friedler

Abstract: Optimal synthesis of workflow structures, the formerly undefined problem, has been introduced. Mathematical programming model is presented for determining the cost optimal workflow system of a given workflow problem. On the basis of a methodology developed for process network synthesis, effective solvers are available for the systematic synthesis of workflow systems.

HTML

XML

PDF

DOI: 10.3217/jucs-012-06-0581

Authors: Benjamin Blankertz, Guido Dornhege, Steven Lemm, Matthias Krauledat, Gabriel Curio, Klaus-Robert Müller

Abstract: We outline the Berlin Brain-Computer Interface (BBCI), a system which enables us to translate brain signals from movements or movement intentions into control commands. The main contribution of the BBCI, which is a non-invasive EEG-based BCI system, is the use of advanced machine learning techniques that allow to adapt to the specific brain signatures of each user with literally no training. In BBCI a calibration session of about 20min is necessary to provide a data basis from which the individualized brain signatures are inferred. This is very much in contrast to conventional BCI approaches that rely on operand conditioning and need extensive subject training of the order 50-100 hours. Our machine learning concept thus allows to achieve high quality feedback already after the very first session. This work reviews a broad range of investigations and experiments that have been performed within the BBCI project. In addition to these general paradigmatic BCI results, this work provides a condensed outline of the underlying machine learning and signal processing techniques that make the BBCI succeed. In the first experimental paradgm we analyze the predictability of limb movement long before the actual movement takes place using only the movement intention measured from the pre-movement (readiness) EEG potentials. The experiments include both off-line studies and an online feedback paradigm. The limits with respect to the spatial resolution of the somatotopy are explored by contrasting brain patterns of movements of left vs. right hand rsp. foot. In a second conplementary paradigm voluntary modulations of sensorimotor rhythms caused by motor imagery (left hand vs. right hand vs. foot) are translated into a continuous feedback signal. Here we report results of a recent feedback study with 6 healthy subjects with no or very little experience with BCI control: half of the subjects achieved an information transfer rate above 35 bits per minute (bmp). Furthermore one subject used the BBCI to operate a mental typewriter in free spelling mode. The overall spelling speed was 4.5-8 letters per minute including the time needed for the correction errors.

HTML

XML

PDF

DOI: 10.3217/jucs-012-04-0370

Authors: Giuseppe Ascia, Vincenzo Catania, Maurizio Palesi

Abstract: Advances in technology now make it possible to integrate hundreds of cores (e.g. general or special purpose processors, embedded memories, application specific components, mixed-signal I/O cores) in a single silicon die. The large number of resources that have to communicate makes the use of interconnection systems based on shared buses inefficient. One way to solve the problem of on-chip communications is to use a Network-on-Chip (NoC)-based communication infrastructure. Such interconnection systems offer new degrees of freedom, exploration of which may reveal significant optimization possibilities: the possibility of arranging the computing and storage resources in an NoC, for example, has a great impact on various performance indexes. The paper addresses the problem of topological mapping of intellectual properties (IPs) on the tiles of a mesh-based NoC architecture. The aim is to obtain the Pareto mappings that maximize performance and minimize power dissipation. We propose a heuristic technique based on evolutionary computing to obtain an optimal approximation of the Pareto-optimal front in an efficient and accurate way. At the same time, two of the most widely-known approaches to mapping in mesh­based NoC architectures are extended in order to explore the mapping space in a multi-criteria mode. The approaches are then evaluated and compared, in terms of both accuracy and efficiency, on a platform based on an event-driven trace-based simulator which makes it possible to take account of important dynamic effects that have a great impact on mapping. The evaluation performed on both synthesized traffic and real applications (an MPEG-4 codec) confirms the efficiency, accuracy and scalability of the proposed approach.

HTML

XML

PDF

DOI: 10.3217/jucs-008-11-0992

Authors: Frédéric Messine

Abstract: Global optimization methods in connection with interval arithmetic permit to determine an accurate enclosure of the global optimum, and of all the corresponding optimizers. One of the main features of these algorithms consists in the construction of an interval function which produces an enclosure of the range of the studied function over a box (right parallelepiped). We use here affine arithmetic in global optimization algorithms, in order to elaborate new inclusion functions. These techniques are implemented and then discussed. Three new affine and quadratic forms are introduced. On some polynomial examples, we show that these new tools often yield more efficient lower bounds (and upper bounds) compared to several well-known classical inclusion functions. The three new methods, presented in this paper, are integrated into various Branch and Bound algorithms. This leads to improve the convergence of these algorithms by attenuating some negative effects due to the use of interval analysis and standard affne arithmetic.

HTML

XML

PDF

DOI: 10.3217/jucs-006-04-0422

Authors: Craig Eldershaw, Stephen Cameron

Abstract: Motion planning is a field of growing importance as more and more computer controlled devices are being used. Many different approaches exist to motion planning|none of them ideal in all situations. This paper considers how to convert a general motion planning problem into one of global optimisation. We regard the general problem as being the classical configuration space findpath problem, but assume that the configurations of the device can be bounded by a hierarchy of hyper-spheres rather than being explicitly computed. A program to solve this problem has been written employing Genetic Algorithms. This paper describes how this was done, and some preliminary results of using it.

HTML

XML

PDF

DOI: 10.3217/jucs-003-07-0821

Authors: Peter Undrill, Kostas Delibasis, George Cameron

Abstract: Stack filters are a class of non-linear spatial operators used for suppression of noise in signals. In this work their design is formulated as an optimisation problem and a method that uses Genetic Algorithms (GAs) to perform the configuration is explained. Because of its computational complexity the process has been implemented as a distributed parallel GA using the Parallel Virtual Machine (PVM) software. We present the results of applying our stack filters to the restoration of magnetic resonance (MR) images corrupted with uniform, uncorellated, noise showing improved statistical performance compared with the median filter and indicating better retention of image details. The efficiency of the parallel implementation is examined, addressing both algorithmic and data decomposition, showing that execution times can be significantly reduced by distributing the task across a network of heterogeneous processors.

HTML

XML

PDF