Optimization of heuristic search using recursive algorithm selection and reinforcement learning

The traditional approach to computational problem solving is to use one of the available algorithms to obtain solutions for all given instances of a problem. However, typically not all instances are the same, nor a single algorithm performs best on all instances. Our work investigates a more sophisticated approach to problem solving, called Recursive Algorithm Selection, whereby several algorithms for a problem (including some recursive ones) are available to an agent that makes an informed decision on which algorithm to select for handling each sub-instance of a problem at each recursive call made while solving an instance. Reinforcement learning methods are used for learning decision policies that optimize any given performance criterion (time, memory, or a combination thereof) from actual execution and profiling experience. This paper focuses on the well-known problem of state-space heuristic search and combines the A* and RBFS algorithms to yield a hybrid search algorithm, whose decision policy is learned using the Least-Squares Policy Iteration (LSPI) algorithm. Our benchmark problem domain involves shortest path finding problems in a real-world dataset encoding the entire street network of the District of Columbia (DC), USA. The derived hybrid algorithm exhibits better performance results than the individual algorithms in the majority of cases according to a variety of performance criteria balancing time and memory. It is noted that the proposed methodology is generic, can be applied to a variety of other problems, and requires no prior knowledge about the individual algorithms used or the properties of the underlying problem instances being solved.     

Autonomous mobile agent routing for efficient server resource allocation

Mobile agents are becoming increasingly important in the highly distributed applications frameworks seen today. Their routing/dispatching from node to node is a very important issue as we need to safeguard application efficiency, achieve better load balancing and resource utilization throughout the underlying network. Selecting the best target server for dispatching a mobile agent is, therefore, a multi-faceted problem that needs to be carefully tackled. In this paper we propose distributed, adaptive routing schemes (next node selection) for mobile agents. The proposed schemes overcome risks like load oscillations, i.e., agents simultaneously abandoning a congested node in search for other, less saturated node. We try to induce different routing decisions taken by agents to achieve load balancing and better utilization of network resources. We consider five different algorithms and evaluate them through simulations. Our findings are quite promising both from the user/application and the network/infrastructure perspective.     

HECTOR: Enabling Microarray Experiments over the Hellenic Grid Infrastructure

Biologists, medical experts, biochemical engineers and researchers working on DNA microarray experiments are increasingly turning on Grid computing with the scope of leveraging the Grid’s computing power, immense storage resources, and quality of service to the expedient processing of a wide range of datasets. In this paper we present a combined experience of grid application experts and bioinformatics scientists in deploying a pilot service enabling computationally efficient processing and analysis of data stemming from microarray experiments. This pilot service is accessible over the Hellenic portion of the EGEE grid and has been demonstrated in the scope of several public events. We highlight the process of grid application enablement, grid deployment challenges, as well as lessons learnt from a bi-annual effort to port and deploy a MATLAB DNA microarray application on a production grid. In addition to describing the parallelization of the application, we also emphasize on the development of a distributed federated database for storing and post-processing the results of the microarray experiments. Overall we believe that our experience could be proven valuable not only to microarray data scientists but also to other Grid users that intend to Grid-enable and deploy their applications.     

Unfolding San Lorenzo

This paper proposes a “reading” of the church of San Lorenzo in Turin, designed by Guarino Guarini, through the philosophical notion of “fold” introduced by Gilles Deleuze. The paper consists of two parts. The first part contains an exploration of the notion of “fold” in architecture and in philosophy and examines the use of the fold in the theory of Baroque architecture as well as the range of this new tool in architectural practise in contemporary architecture and in philosophy and examines the use of the fold as fundamental condition for understanding Baroque era. The second part contains the application of the notion of fold as a philosophical and conceptual framework for the “reading” of the chapel.     

Flow on the Internet: a longitudinal study of Internet addiction symptoms during adolescence

Internet Addiction (IA) constitutes an excessive Internet use behavior with a significant impact on the user’s well-being. Online flow describes the users’ level of being absorbed by their online activity. The present study investigated age-related, gender, and flow effects on IA in adolescence. The sample comprised 648 adolescents who were assessed twice at age 16 and 18 years. IA was assessed using the Internet Addiction Test and online flow was assessed using the Online Flow Questionnaire. A three-level hierarchical model estimated age-related, gender, and online flow effects on IA symptoms and controlled for clustered random effects. IA symptoms decreased over time (for both genders) with a slower rate in males. Online flow was associated with IA symptoms and this remained consistent over time. Findings expand upon the available literature suggesting that IA symptoms could function as a development-related manifestation at the age of 16 years, while IA-related gender differences gradually increase between 16 and 18 years. Finally, the association between online flow and IA symptoms remained stable independent of age-related effects. The study highlights individual differences and provides directions for more targeted prevention and intervention initiatives for IA.     

Improved transmission line model for high-frequency modelling of through silicon vias

Accurate and reliable models can support Through Silicon Via (TSV) testing methods and improve the quality of 3D ICs. A model for expressing resistance and inductance of TSVs at frequencies up to 50 GHz is proposed. It is based on the two-parallel transmission cylindrical wires model, known also as the Transmission Line Model and improved through the fitting to ANSYS Q3D simulation results. The proximity effect between neighbouring TSVs that alters the paths through which current flows is empowered at high frequencies. The consideration of the dependence of the proximity effect on frequency for calculating TSV resistance and inductance is the main contribution of this work. Additionally, the modelling of resistance is extended to accurately correspond to a TSV in an array. The proposed models are in good agreement with the simulator results with an average error below 2% and 5.4% for the resistance and the inductance, respectively. The maximum error is 3% and 9.1%, respectively. In the case of the resistance of a TSV in an array, the maximum error is 4.7%. As long as the coefficients of the proposed equations have been extracted, the time for resistance and inductance calculation based on the presented models is negligible, compared to the time-consuming EM simulation.     

A data locality methodology for matrix–matrix multiplication algorithm

Matrix-Matrix Multiplication (MMM) is a highly important kernel in linear algebra algorithms and the performance of its implementations depends on the memory utilization and data locality. There are MMM algorithms, such as standard, Strassen–Winograd variant, and many recursive array layouts, such as Z-Morton or U-Morton. However, their data locality is lower than that of the proposed methodology. Moreover, several SOA (state of the art) self-tuning libraries exist, such as ATLAS for MMM algorithm, which tests many MMM implementations. During the installation of ATLAS, on the one hand an extremely complex empirical tuning step is required, and on the other hand a large number of compiler options are used, both of which are not included in the scope of this paper. In this paper, a new methodology using the standard MMM algorithm is presented, achieving improved performance by focusing on data locality (both temporal and spatial). This methodology finds the scheduling which conforms with the optimum memory management. Compared with (Chatterjee et al. in IEEE Trans. Parallel Distrib. Syst. 13:1105, 2002; Li and Garzaran in Proc. of Lang. Compil. Parallel Comput., 2005; Bilmes et al. in Proc. of the 11th ACM Int. Conf. Super-comput., 1997; Aberdeen and Baxter in Concurr. Comput. Pract. Exp. 13:103, 2001), the proposed methodology has two major advantages. Firstly, the scheduling used for the tile level is different from the element level’s one, having better data locality, suited to the sizes of memory hierarchy. Secondly, its exploration time is short, because it searches only for the number of the level of tiling used, and between (1, 2) (Sect. 4) for finding the best tile size for each cache level. A software tool (C-code) implementing the above methodology was developed, having the hardware model and the matrix sizes as input. This methodology has better performance against others at a wide range of architectures. Compared with the best existing related work, which we implemented, better performance up to 55% than the Standard MMM algorithm and up to 35% than Strassen’s is observed, both under recursive data array layouts.     

Effect of replacing soybean meal by extruded chickpeas in the diets of growing-finishing pigs on meat quality

The main objective of this study was to evaluate the effect of the replacement of soybean meal by extruded chickpeas in diets of growing-finishing pigs on meat quality. In a 17wk study 48 growing-finishing crossbred pigs were fed ad libitum. The experimental design included four treatments, each one of 12 pigs; the ECKP0 treatment was fed with diet containing soybean meal and no chickpeas (control), while treatments ECKP100, ECKP200 and ECKP300 were fed with diets containing 100, 200 and 300kg/t of extruded chickpeas, respectively. The lean meat quality of the longissimus lumborum et thoracis muscle was evaluated by chemical analysis (moisture, protein, fat and ash), fatty acid profile, pH measurement, cooking loss, color evaluation, and sensory evaluation. Odor and taste, tenderness, juiciness, and overall acceptability were scored on 1-10 scales by a group of 10 experienced assessors after a standard cooking regime. Small differences were observed between control and experimental groups in chemical composition (P>0.05). Fatty acid profiles, pH measurements and color evaluation did not differ among treatments (P>0.05), while cooking loss was significantly lower in the control group (P<0.05). The taste panel gave slightly higher scores for the tenderness and juiciness for the control group compared with the chickpea treatments (P<0.05). No differences were observed between control and experimental groups in taste scores (P>0.05). It is concluded that the replacement of soybean meal by extruded chickpeas, when substituted isonitrogenously and isoenergetically at inclusion levels up to 300kg/t of pig, does not influence significantly meat quality.     

Cost analysis and efficient radio bearer selection for multicasting over UMTS

Along with the widespread deployment of the Third Generation (3G) cellular networks, the fast‐improving capabilities of the mobile devices, content, and service providers are increasingly interested in supporting multicast communications over wireless networks and in particular over Universal Mobile Telecommunications System (UMTS). To this direction, the Third Generation Partnership Project (3GPP) is currently standardizing the Multimedia Broadcast/Multicast Service (MBMS) framework of UMTS. In this paper, we present an overview of the MBMS multicast mode of UMTS. We analytically present the multicast mode of the MBMS and analyze its performance in terms of packet delivery cost under various network topologies, cell types, and multicast users' distributions. Furthermore, for the evaluation of the scheme, we consider different transport channels for the transmission of the multicast data over the UMTS Terrestrial Radio‐Access Network (UTRAN) interfaces. Finally, we propose a scheme for the efficient radio bearer selection that minimizes total packet delivery cost. Copyright © 2008 John Wiley & Sons, Ltd.     

On the efficient simulation of electrical circuits with constant phase elements: The Warburg element as a test case

The constant phase Element (CPE) concept naturally emerges as a model for describing a range of electrical phenomena where ionic diffusion is involved. We suggest a new method for modelling the transient behavior of electrical circuits that contain CPE elements. Without loss of generality, we study the Warburg element to demonstrate the method, but the method can be easily extended to any CPE. Transient simulations of such elements require the numerical evaluation of a computationally expensive convolution integral that links the voltage drop across the element, with the current that passed through it. In our work we suggest a new method for reducing the computational cost of the numerical evaluation of the convolution integral. We show that the computational cost can be reduced by one order of magnitude.