Session2 15:45 - 17:25 on 12th June 2017

ICCS 2017 Main Track (MT) Session 2

Time and Date: 15:45 - 17:25 on 12th June 2017

Room: HG F 30

Chair: Anna-Lena Lamprecht

341 Resolving Entity Morphs based on Character-Word Embedding [abstract]
Abstract: Morph is a special type of fake alternative names. Internet users use morphs to achieve certain goals such as expressing special sentiment or avoiding censorship. For example, Chinese internet users often replace “马景涛” (Ma Jingtao) with “咆哮教主”(Roar Bishop). “咆哮教主”(Roar Bishop) is a morph and “马景涛” (Ma Jingtao) is the target entity of “咆哮教主”(Roar Bishop). This paper mainly focuses on morph resolution: given a morph, figure out the entity that it really refers to. After analysis the common characteristic of morphs and target entities from cross-source corpora, we exploit temporal and semantic constraints to collect target candidates. Next, we propose a character-word embeddings framework to rank target candidates. Our method does not need any human-annotated data. Experimental results demonstrate that our approach outperforms the state-of-the-art method. The results also show that the performance is better when morphs share any character with target entities.
Ying Sha, Zhenhui Shi, Rui Li, Qi Liang, Bin Wang and Li Guo
273 Graph Ranking Guarantees for Numerical Approximations to Katz Centrality [abstract]
Abstract: Graphs and networks are prevalent in modeling relational datasets from many fields of research. Using iterative solvers to approximate graph measures (specifically Katz Centrality) allows us to obtain a ranking vector, consisting of a number for each vertex in the graph identifying its relative importance. We use the residual to accurately estimate how much of the ranking from an approximate solution matches the ranking given by the exact solution. Using probabilistic matrix norms and applying numerical analysis to the computation of Katz Centrality, we obtain bounds on the accuracy of the approximation compared to the exact solution with respect to the highly ranked nodes. This relates the numerical accuracy of the linear solver to the data analysis accuracy of finding the correct ranking. In particular, we answer the question of which pairwise rankings are reliable given an approximate solution to the linear system. Experiments on many real-world networks up to several million vertices and several hundred million edges validate our theory and show that we are able to accurately estimate large portions of the approximation. By analyzing convergence error, we develop confidence in the ranking schemes of data mining.
Eisha Nathan, Geoffrey Sanders, James Fairbanks, Van Henson and David Bader
447 Simulating a Search Engine Service focusing on Network Performance [abstract]
Abstract: Large-scale computer systems like Search Engines provide services to thousands of users, and their user demand can change suddenly. This unstable demand impacts sensitively to the service components (like network and hosts). The system should be able to address unexpected scenarios; otherwise, users would be forced to leave the service. Creating tests scenarios is an alternative to deal with this variable workload before implementing new configuration in the system. However, the complexity and size of the system are a huge constraint to create physical models. Simulation can help to test promising models of search engines. In this paper we propose a method to model a Search Engine Service (SES) on small scale to analyze the impact of different configurations. We model the interaction of a typical search engine with three main components: a Front Service (FS), a Cache Service (CS) and an Index service (IS). The FS takes as input a query of user and search into a database with the support of a CS to improve the performance of the system. The proposed model processes a trace file from a real SES and based on the dependency relation among the messages, services and queries, it is modeled the full functionality of the SES. The output is, on one hand a simulated trace file to compare the model with the real system and on the other hand statistics about performance. The simulation allow us to test configurations of FS, CS, and IS, which can be unlikely in the real system.
Joe Carrión, Daniel Franco and Emilo Luque
245 Fully-Dynamic Graph Algorithms with Sublinear Time Inspired by Distributed Computing [abstract]
Abstract: We study dynamic graphs in the fully-dynamic centralized setting. In this setting the vertex set of size n of a graph G is fixed, and the edge set changes step-by-step, such that each step either adds or removes an edge. The goal in this setting is maintaining a solution to a certain problem (e.g., maximal matching, edge coloring) after each step, such that each step is executed efficiently. The running time of a step is called update-time. One can think of this setting as a dynamic network that is monitored by a central processor that is responsible for maintaining the solution. Currently, for several central problems, the best-known deterministic algorithms for general graphs are the naive ones which have update-time O(n). This is the case for maximal matching and proper O(Delta)-edge-coloring. The question of existence of sublinear in n update-time deterministic algorithms for dense graphs and general graphs remained wide open. In this paper we address this question by devising sublinear update-time deterministic algorithms for maximal matching in graphs with bounded neighborhood independence o(n/ log^2 n), and for proper O(Delta)-edge-coloring in general graphs. The family of graphs with bounded neighborhood independence is a very wide family of dense graphs. In particular, graphs with constant neighborhood independence include line-graphs, claw-free graphs, unit disk graphs, and many other graphs. Thus, these graphs represent very well various types of networks. For graphs with constant neighborhood independence, our maximal matching algorithm has ~O(\sqrt n) update-time. Our O(Delta)-edge-coloring algorithms has ~O(\sqrt Delta) update-time for general graphs. In order to obtain our results we employ a novel approach that adapts certain distributed algorithms of the LOCAL setting to the centralized fully-dynamic setting. This is achieved by optimizing the work each processors performs, and efficiently simulating a distributed algorithm in a centralized setting. The simulation is efficient thanks to a careful selection of the network parts that the algorithm is invoked on, and by deducing the solution from the additional information that is present in the centralized setting, but not in the distributed one. Our experiments on various network topologies and scenarios demonstrate that our algorithms are highly-efficient in practice. We believe that our approach is of independent interest and may be applicable to additional problems.
Leonid Barenboim and Tzalik Maimon

ICCS 2017 Main Track (MT) Session 9

Time and Date: 15:45 - 17:25 on 12th June 2017

Room: HG D 1.1

Chair: Craig Douglas

343 An Ensemble of Kernel Ridge Regression for Multi-class Classification [abstract]
Abstract: We propose an ensemble of kernel ridge regression based classifiers in this paper. Kernel ridge regression admits a closed form solution making it faster to compute and also making it suitable to use for ensemble methods for small and medium sized data sets. Our method uses random vector functional link network to generate training samples for kernel ridge regression classifiers. Several kernel ridge regression classifiers are constructed from different training subsets in each base classifier. The partitioning of the training samples into different subsets leads to a reduction in computational complexity when calculating matrix inverse compared with the standard approach of using all N samples for kernel matrix inversion. The proposed method is evaluated using well known multi-class UCI data sets. Experimental results show the proposed ensemble method outperforms the single kernel ridge regression classifier and its bagging version.
Rakesh Katuwal and Ponnuthurai Suganthan
385 Dynamic Profiles Using Sentiment Analysis for VAA’s Recommendation Design [abstract]
Abstract: In the context of elections, the Internet opens new and promising possibilities for parties and candidates looking for a better political strategy and visibility. In this way they can also organize their election campaign to gather funds, to mobilize support, and to enter into a direct dialogue with the electorate. This paper presents an ongoing research of recommender systems applied on e-government, particularly it is an extension of so-called voting advice applications (VAA's). VAA's are Web applications that support voters, providing relevant information on candidates and political parties by comparing their political interests with parties or candidates on different political issues. Traditional VAA's provide recommendations of political parties and candidates focusing on static profiles of users. The goal of this work is to develop a candidate profile based on different parameters, such as the perspective of voters, social network activities, and expert opinions, to construct a more accurate dynamic profile of candidates. Understanding the elements that compose a candidate profile will help citizens in the decision-making process when facing a lack of information related to the behavior and thinking of future public authorities. At the end of this work, a fuzzy-based visualization approach for a VAA design is given using as a case study the National Elections of Ecuador in 2013.
Luis Terán and Jose Mancera
25 Discriminative Learning from Selective Recommendation and Its Application in AdaBoost [abstract]
Abstract: The integration of semi-supervised learning and ensemble learning has been a promising research area. It is a typical procedure that one learner recommends the pseudo-labeled instances with high predictive confidence to another, so that the training dataset is expanded. However, the new learner’s demand on recommendation as well as the possibility of incorrect recommendation are neglected, which inevitably jeopardize the learning performance. To address these issues, this paper proposes the Discriminative Learning from Selective Recommendation (DLSR) method. On one hand, both reliability and informativeness of the pseudo-labeled instances are taken into account via selective recommendation. On the other hand, the potential in both correct and incorrect recommendation are formulated in discriminative learning. Based on DLSR, we further propose the selective semi-supervised AdaBoost. With both recommending and receiving learners engaged in ensemble model learning, the unlabeled instances are explored in a more effective way.
Xiao-Yu Zhang, Shupeng Wang, Chao Li, Shiming Ge, Yong Wang and Binbin Li
157 Distributed Automatic Differentiation for Ptychography [abstract]
Abstract: Synchrotron radiation light source facilities are leading the way to ultrahigh resolution X-ray imaging. High resolution imaging is essential to understanding the fundamental structure and interaction of materials at the smallest length scale possible. Diffraction based methods achieve nanoscale imaging by replacing traditional objective lenses by pixelated area detectors and computational image reconstruction. Among these methods, ptychography is quickly becoming the standard for sub-30 nanometer imaging of extended samples, but at the expense of increasingly high data rates and volumes. This paper presents a new distributed algorithm for solving the ptychographic image reconstruction problem based on automatic differentiation. Input datasets are subdivided between multiple graphics processing units (GPUs); each subset of the problem is then solved either entirely independent of other subsets (asynchronously) or through sharing gradient information with other GPUs (synchronously). The algorithm was evaluated on simulated and real data acquired at the Advanced Photon Source, scaling up to 192 GPUs. The synchronous variant of our method outperformed an existing multi-GPU implementation in terms of accuracy while running at a comparable execution time.
Youssef Nashed, Tom Peterka, Junjing Deng and Chris Jacobsen
57 Automatic Segmentation of Chinese Characters as Wire-Frame Models [abstract]
Abstract: There exist thousands of Chinese characters, used across several countries and languages. Their huge number induces various processing difficulties by computers. One challenging topic is for example the automatic font generation for such characters. Also, as these characters are in many cases recursive compounds, pattern (i.e. sub-character) detection is an insightful topic. In this paper, aiming at addressing such issues, we describe a segmentation method for Chinese characters, producing wire-frame models, thus vector graphics, compared to conventional raster approaches. While raster output would enable only very limited reusing of these wire-frame models, vector output would for instance support the automatic generation of vector fonts (Adobe Type 1, Apple True Type, etc.) for such characters. Our approach also enables significant performance increase compared to the raster approach. The proposed method is then experimented with a list of several Chinese characters. Next, the method is empirically evaluated and its average time complexity is assessed.
Antoine Bossard

ICCS 2017 Main Track (MT) Session 16

Time and Date: 15:45 - 17:25 on 12th June 2017

Room: HG D 1.2

Chair: Fabrício Enembreck

135 StoreRush: An Application-Level Approach to Harvesting Idle Storage in a Best Effort Environment [abstract]
Abstract: For a production HPC system where storage devices are shared between multiple applica- tions and managed in a best effort manner, contention is often a major problem leading to some storage devices being more loaded than others and causing a significant reduction in I/O throughput. In this paper, we describe our latest efforts StoreRush to resolve this practical issue at the application level without requiring modification to the file and storage system. The proposed scheme uses a two-level messaging system to harvest idle storage via re-routing I/O requests to a less congested storage location so that write performance is improved while lim- iting the impact on read by throttling re-routing if deemed too much. An analytical model is derived to guide the setup of optimal throttling factor. The proposed scheme is verified against production applications Pixie3D, XGC1 and QMCPack during production windows, which very well demonstrated the effectiveness (e.g., up to 1.8x improvement in write) and scalability of our approach (up to 131,072 cores).
Qing Liu, Norbert Podhorszki, Jong Choi, Jeremy Logan, Matt Wolf, Scott Klasky, Tahsin Kurc and Xubin He
204 Fast Parallel Construction of Correlation Similarity Matrices for Gene Co-Expression Networks on Multicore Clusters [abstract]
Abstract: Gene co-expression networks are gaining attention in the present days as useful representations of biologically interesting interactions among genes. The most computationally demanding step to generate these networks is the construction of the correlation similarity matrix, as all pairwise combinations must be analyzed and complexity increases quadratically with the number of genes. In this paper we present MPICorMat, a hybrid MPI/OpenMP parallel approach to construct similarity matrices based on Pearson’s correlation. It is based on a previous tool (RMTGeneNet) that has been used on several biological studies and proved accurate. Our tool obtains the same results as RMTGeneNet but significantly reduces runtime on multicore clusters. For instance, MPICorMat generates the correlation matrix of a dataset with 61,170 genes and 160 samples in less than one minute using 16 nodes with two Intel Xeon Sandy-Bridge processors each (256 total cores), while the original tool needed almost 4.5 hours. The tool is also compared to another available approach to construct correlation matrices on multicore clusters, showing better scalability and performance. MPICorMat is an open-source software and it is publicly available at https://sourceforge.net/projects/mpicormat/.
Jorge González-Domínguez and María J. Martín
261 The Design and Performance of Batched BLAS on Modern High-Performance Computing Systems [abstract]
Abstract: A current trend in high-performance computing is to decompose a large linear algebra problem into batches containing thousands of smaller problems, that can be solved independently, before collating the results. To standardize the interface to these routines, the community is developing an extension to the BLAS standard (the batched BLAS), enabling users to perform thousands of small BLAS operations in parallel whilst making efficient use of their hardware. We discuss the benefits and drawbacks of the current batched BLAS proposals and perform a number of experiments, focusing on GEMM, to explore their affect on the performance. In particular we analyze the effect of novel data layouts which, for example, interleave the matrices in memory to aid vectorization and prefetching of data. Utilizing these modifications our code outperforms both MKL and CuBLAS by up to 6 times on the self-hosted Intel KNL (codenamed Knights Landing) and Kepler GPU architectures, respectively, for large numbers of DGEMM operations using matrices of size 2 × 2 to 20 × 20
Jack Dongarra, Sven Hammarling, Nick Higham, Samuel Relton, Pedro Valero-Lara and Mawussi Zounon
333 OUTRIDER: Optimizing the mUtation Testing pRocess In Distributed EnviRonments [abstract]
Abstract: The adoption of commodity clusters has been widely extended due to its cost-effectiveness and the evolution of networks. These systems can be used to reduce the long execution time of applications that require a vast amount of computational resources, and especially of those techniques that are usually deployed in centralized environments, like testing. Currently, one of the main challenges in testing is to obtain an appropriate test suite. Mutation testing is a widely used technique aimed at generating high quality test suites. However, the execution of this technique requires a high computational cost. In this work we propose OUTRIDER, an HPC-based optimization that contributes to bridging the gap between the high computational cost of mutation testing and the parallel infrastructures of HPC systems aimed to speed-up the execution of computational applications. This optimization is based on our previous work called EMINENT, an algorithm focused on parallelizing the mutation testing process using MPI. However, since EMINENT does not efficiently exploit the computational resources in HPC systems, we propose 4 strategies to alleviate this issue. A thorough experimental study using different applications shows an increase of up to 70% performance improvement using these optimizations.
Pablo C. Cañizares, Alberto Núñez and Juan de Lara
112 Topology-aware Job Allocation in 3D Torus-based HPC Systems with Hard Job Priority Constraints [abstract]
Abstract: In this paper, we address the topology-aware job allocation problem on 3D torus-based high performance computing systems, with the objective of reducing system fragmentation. Firstly, we propose a group-based job allocation strategy, which leads to a more global optimization of resource allocation. Secondly, we propose two shape allocation methods to determine the topological shape for each input job, including a zigzag allocation method for communication non-sensitive jobs, and a convex allocation method for communication sensitive jobs. Thirdly, we propose a topology-aware job mapping algorithm to reduce the system fragmentation brought in by the job mapping process, including a target bin selection method and a bi-directional job mapping method. The evaluation results validate the efficiency of our approach in reducing system fragmentation and improving system utilization.
Kangkang Li, Maciej Malawski and Jarek Nabrzyski

Agent-based simulations, adaptive algorithms and solvers (ABS-AAS) Session 2

Time and Date: 15:45 - 17:25 on 12th June 2017

Room: HG D 7.1

Chair: Quenling Deng

156 A wrapper around parallel MUMPS solver to reduce its memory usage and execution time for finite element method computations [abstract]
Abstract: In this paper, we present a wrapper around MUMPS solver, called Hierarchical Solver Wrapper (HSW) that allows reducing its memory usage and execution time. The wrapper is dedicated for domain decomposition based parallel finite element method computations on Linux cluster. It utilizes the identical interface as parallel MUMPS with entries provided as lists of non-zero entries distributed among multiple processors. It wraps around MUMPS solver, performing two level hierarchical calls. First, it calls multiple sequential MUMPS solvers, one per each processor. They compute Schur complements of the interior nodes on the interface nodes over each subdomain. It deallocates the sequential MUMPS solvers on every processor to minimize the memory usage. It keeps only original lists of non-zero entries as well as lists of non-zero entries obtained from the computed Schur complement. It submits all the Schur complements to one parallel MUMPS solver and asks for the skeleton problem solution. Later, it restores the upper part of the local matrix to perform local backward substitutions with Schur complements replaced by the identity matrices and right-hand-sides superseded by the local part of the global solution. The wrapper is tested with latest version of MUMPS over the three-dimensional isogeometric analysis computations. It enables for reduction of the memory usage and also the execution time, in comparison with a single parallel MUMPS call with distributed lists of non-zero entries.
Maciej Paszynski and Antonio Tadeu Gomes
48 Non-Fitting meshes for Maxwell's equations [abstract]
Abstract: In context of adaptive finite element methods, we explore the possibility of employing lowest-order non-fitting meshes for solving Maxwell´s equations. In a fitting mesh, the physical interfaces of the propagation medium must be aligned with cell faces. On the other hand, this constraint is removed for non-fitting meshes, so that a larger choice of cells is possible. As a result, non-fitting meshes can simplify the implementation and/or reduce the computational cost of the associated finite-element method. Unfortunately, non-fitting meshes can also cause an accuracy loss. Indeed, the solution of Maxwell's equations can become singular inside mesh cells due to material discontinuities. In this work, we carefully analyze the accuracy loss due to the use of non-fitting meshes. We derive error estimates that are further confirmed via numerical experiments. The main conclusion is that electric and magnetic fields exhibit different convergence rates in the L^2-norm for the case of non-fitting meshes. In particular, if the user is interested in the magnetic field, non-fitting meshes produce solutions of similar quality to those obtained using fitting meshes.
Théophile Chaumont-Frelet and David Pardo
51 Fast Isogeometric L2 Projection Solver for Tumor Growth Simulations [abstract]
Abstract: In this talk, we focus on the application of the isogeometric analysis for tumor growth simulations. We present an application of the fast algorithm for isogeometric L2 projections for simulations of the tumor growth. We utilize the of PDE describing the model of the melanoma growth, including tumor cell density, flux, pressure, extracellular and degraded extracellular matrices, previously used in finite difference simulations. We also use a discrete model of the vasculature that provides an oxygen source to the system. The system is solved using explicit scheme and fast isogeometric L2 projections, utilizing the alternating directions solver. Every 10-time steps of the simulation we couple our continuous model with the discrete vasculature model. The presentation is concluded with the two-dimensional numerical results. We show that explicit dynamics simulation needs around 30,000 times steps of the alternating direction solver to be performed in two-dimensions.
Maciej Paszynski, Witold Dzwinel, Marcin Łoś and Adrian Kłusek
86 Goal-Oriented p-Adaptivity using Unconventional Error Representations for a 1D Steady State Convection-Diffusion Problem [abstract]
Abstract: This work proposes the use of an alternative error representation for Goal-Oriented Adaptivity (GOA) in context of steady state convection dominated diffusion problems. It introduces an arbitrary operator for the computation of the error of an alternative adjoint problem. From the new representation, we derive element-wise estimators to drive the adaptive algorithm. The method is applied to a one dimensional (1D) steady state convection dominated diffusion problem with homogeneous Dirichlet boundary conditions. This problem exhibits a boundary layer that produces a loss of numerical stability. The new error representation delivers sharper error bounds. When applied to a p-GOA Finite Element Method (FEM), the alternative error representation captures earlier the boundary layer, despite the existing spurious numerical oscillations.
Vincent Darrigrand, Ángel Rodríguez-Rozas, David Pardo and Ignacio Muga
224 Algorithms for construction of Element Partition Trees for Direct Solver executed over h refined grids with B-splines and C0 separators [abstract]
Abstract: We propose a way of performing isogeometric finite element method (IGA-FEM) computations over h refined grids with B-spline basis functions. Namely, we propose to use the B-spline basis functions defined over patches of elements with C0 separators between the refinement levels. Our definition of the B-splines and C0 separators allows introduction of arbitrary order B-splines over 2D grids refined towards singularities. We also present an algorithm for construction of element partition trees (EPT) over h refined grids with modified B-splines. The EPT allows generating an ordering which gives a linear computational cost of the multi-frontal solver over 2D grids refined towards a point or an edge singularity. We present the algorithm for transforming the EPT into an ordering. We also verify the linear computational cost of the proposed method on grid with point and edge singularity. We compare our method to h-adaptive finite element method (h-FEM) computations with Lagrange polynomials.
Bartosz Janota and Maciej Paszynski

Simulations of Flow and Transport: Modeling, Algorithms and Computation (SOFTMAC) Session 2

Time and Date: 15:45 - 17:25 on 12th June 2017

Room: HG D 7.2

Chair: Shuyu Sun

196 The THex Algorithm and a Simple Darcy Solver on Hexahedral Meshes [abstract]
Abstract: In this paper, we first present the THex algorithm that refines a tetrahedral mesh into a hexahedral mesh. Strategies for efficient implementation of the THex algorithm are discussed. Then we present the lowest order weak Galerkin (WG) $ (Q_0,Q_0;RT_{[0]}) $ finite element method for solving the Darcy equation on general hexahedral meshes. This simple solver uses constant pressure unknowns inside hexahedra and on faces but specifies the discrete weak gradients of these basis functions in local Raviart-Thomas $ RT_{[0]} $ spaces. The solver has easy implementation, is locally mass-conservative, and produces continuous normal fluxes, regardless of hexahedral mesh quality. When the mesh is asymptotically parallelopiped, this Darcy solver exhibits optimal order convergence in pressure, velocity, and flux, as demonstrated by numerical results.
Graham Harper, Jiangguo Liu and Bin Zheng
67 Mixed finite element analysis for an elliptic/mixed elliptic interface problem with jump coefficients [abstract]
Abstract: Aiming at the development of a practically parallelizable algorithm for the simulation of fluid-structure interaction (FSI) problems in the future, in this paper a type of elliptic interface problem with jump coefficients is chosen to begin with and is reformulated to a mixture of elliptic/mixed elliptic interface problem which is analogous to a steady state FSI problem to some extent. A mixture of standard- and mixed finite element method is developed for the reformulation of the elliptic interface problem, and its well-posedness and convergence are studied by proving the inf-sup condition of a total bilinear form. With the body-fitted meshes and a strongly coupled alternating iteration scheme, numerical experiments are carried out for an elliptic interface problem with an immersed interface for different jump ratios, and the obtained numerical results confirm our convergence theorem.
Rihui Lan, Pengtao Sun and Mo Mu
179 Stabilized finite element methods for flux [abstract]
Abstract: In this paper, stabilized continuous finite element methods are analyzed for numerically solving the flux which may be a non $H^1$ solution. Coercivity and error estimates are established. Numerical experiments are performed to illustrate these methods.
Huoyuan Duan
197 Comparison of Handling Pressure in Poisson Solver for Immersed Boundary Method Considering Pressure Condition [abstract]
Abstract: In the Cartesian grid approach, the immersed boundary method (IBM) is well used to handle the boundary of an object with complicated shape on the Cartesian grid. However, the conventional IBM generates the unphysical pressure oscillations near the boundary because of the pressure jump between inside and outside of the boundary. The IBM considering pressure condition was proposed in order to remove the pressure oscillations by solving the governing equations considering the pressure condition on the boundary. In this method, there are two ways of the handling the pressure on the boundary in the Poisson solver. In this paper, the effect of removing the pressure oscillations by the IBM considering the pressure condition is investigated. And, the influence by the difference in the handling of the pressure on the boundary in the Poisson solver is investigated. In the numerical simulations of incompressible flow around a 2D circular cylinder, the present IBM indicate a greate effect of removing the pressure oscillations. And, it do not occur difference of the result by the difference of the handling the pressure on the boundary in the Poisson solver. Therefore, it is possible to select a method with less computational amount in the Poisson solver without degrading the quality of the result. It is concluded that the present IBM is very promising as improved method in order to remove the pressure oscillations in the conventional IBM.
Kyohei Tajiri, Hidetoshi Nishida and Mitsuru Tanaka
350 Inviscid regularization of compressible two-phase flow using observable divergence theorem [abstract]
Abstract: Many fluid flow problems involving turbulence, shocks, and material interfaces create a common issue that we call $k_\infty$ irregularity. The non-linear advection term in the governing equations for all of these problems keep generating higher wave modes as $k$ goes to infinity. In this work, we present an inviscid regularization technique, called observable regularization, for the simulation of two-phase compressible flow. In this technique, we use observable divergence theorem to derive an observable equation for tracking material interface (volume fraction). Using a couple of one-dimensional test cases, first we show that this method preserves pressure equilibrium at material interface, then we compare our results to exact Euler solutions. At the end we demonstrate a two-dimensional simulation of shock-bubble interaction showing good agreement with available experimental data from literature.
Bahman Aboulhasanzadeh and Kamran Mohseni
446 Computational modeling of flow and solute transport in a nephron [abstract]
Abstract: The kidneys are vital organs that contribute to the maintenance of homeostasis in our body. They fulfill functions such as electrolyte control, blood filtration and initiation of red blood cell production in response to hypoxia, a state characterized by deficiency of oxygen (O2 ) in the renal tissue. A normal human kidney contains between 0.8 to 1.5 million functional units called nephrons. Renal tubules along the nephrons are responsible for the reabsorption of various solutes including sodium ions (Na+ ), a process that requires large amounts of O2. Physiological and pathophysiological variation in Na+ transport can alter O2 consumption, leading to changes in tissue oxygenation. We have developed a one-dimensional (1D) mathematical model for the computation of Na+ reabsorption and corresponding O2 consumption along a nephron, which is parameterized using published data obtained from a rat kidney. Our computations demonstrate that per kidney 8μ-moles of O2 are consumed per minute for the reabsorption of 125μ-moles of Na+ , which is in agreement with previously published results. The model also predicts that changes in arterial pressure adversely impact the efficiency of O2 consumption in the nephron. The model is further being extended to account for anatomically realistic renal vasculature obtained from synchrotron X-ray phase contrast micro computed tomography, with the final goal of determining O2 distribution in the whole kidney.
Kartik Jain and Vartan Kurtcuoglu

Multiscale Modelling and Simulation (MMS) Session 2

Time and Date: 15:45 - 17:25 on 12th June 2017

Room: HG D 3.2

Chair: Derek Groen

20 A concept of a prognostic system for personalized anti-tumor therapy based on supermodeling [abstract]
Abstract: Application of computer simulation for predicting cancer progression/remission/recurrence is still underestimated by clinicians. This is mainly due to the lack of tumor modeling approaches, which are both reliable and realistic computationally. We propose and describe here the concept of a viable prediction/correction system for predicting cancer dynamics, very similar, in spirit, to that used in weather forecast and climate modeling. It is based on the supermodeling technique where the supermodel consists of a few coupled instances (sub-models) of a generic coarse-grained tumor model. Consequently, the latent and fine-grained cancer properties not included in the generic model, e.g. reflecting microscopic phenomena and other unpredictable events influencing tumor dynamics, are hidden in sub-models coupling parameters, which can be learned from incoming real data. Thus instead of matching hundreds of parameters for multi-scale tumor models by using complicated scales-bridging and data adaptation schemes, we need to fit only several values of coupling coefficients between sub-models to the current tumor status. Here, we propose a supermodel based, prediction/correction scheme that can be further employed for planning anti-cancer therapy and drug treatment, being continually updated by incoming diagnostic data.
Witold Dzwinel, Adrian Kłusek and Maciej Paszynski
219 Linking Gene Dynamics to Intimal Hyperplasia – A Predictive Model of Vein Graft Adaptation [abstract]
Abstract: The long term outcome of Coronary Artery Bypass Graft (CABG) surgery remains unsatisfactory to this day. Despite years of improvements in surgical techniques and therapies administered, re-occlusion of the graft is experienced in 10-12% of the cases within just few months (Motwani JC, 1998). We suggest that an efficient post-surgical therapy might be found at the genetic level. Accordingly, we propose a multiscale model that is able to replicate the healing of the graft and detail the level of impact of targeted clusters of genes on the graft’s healing. A key feature of our integrated model is its capability of linking the genetic, cellular and tissue levels with feedback bridges in such a way that every single variation from an equilibrium point is reflected on all the other elements, creating a highly organized loop. Once validated on experimental data, our model offers the possibility to test several gene therapies that aim to improve the patency of the graft lumen in advance. Being able to anticipate the outcome will speed up the development of an efficient therapy and may lead to prolonged life expectancy of the graft.
Stefano Casarin, Scott A. Berceli and Marc Garbey
576 Multiscale Computing and Systems Medicine in COST: a Brief Reflection [abstract]
Abstract: Today's modelling approaches in systems medicine are increasingly multiscale, containing two or more submodels, each of which operates on different temporal and/or spatial scales (Hunter,2008). In addition, as these models become increasingly sophisticated, they tend to be run as multiscale computing applications using computational infrastructures such as clusters, supercomputers, grids or clouds. Constructing, validating and deploying such applications is far from trivial, and communities in different scientific disciplines have chosen very diverse approaches to address these challenges (Groen,2014;Borgdorff,2013). Within this presentation we reflect on the use of Multiscale Computing within the context of Open Multiscale Systems Medicine (OpenMultiMed) COST action and related developments. Multiscale computing is widely applied within this area, and instead of summarizing the field as a whole we will highlight a set of challenges that we believe are of key relevance to the systems medicine community. Among these we will highlight key multiscale computing challenges in the context of healthcare and assisted living settings, systems medicine data analytics, effectively exploiting cloud and HPC infrastructures, and the use of Multiscale Computing in relation to the Internet of Things. Note: This abstract is part of a paper-in-progress project by Multiscale Computing Working Group in the OpenMultiMed project, where we reflect on current advances and seek to formulate a vision on Multiscale Computing specific to this COST project. We appreciate any points raised by the referees or during the presentation, and intend to take those to heart for our future activities. (Hunter:2008) Peter J Hunter, Edmund J Crampin, and Poul MF Nielsen. Bioinformatics, multiscale modeling and the iups physiome project. Briefings in bioinformatics, 9(4):333–343, 2008. (Groen:2014) Derek Groen, Stefan J Zasada, and Peter V Coveney. Survey of multiscale and multiphysics applications and communities. Computing in Science & Engineering, 16(2):34–43, 2014. (Borgdorff:2013) Joris Borgdorff, Jean-Luc Falcone, Eric Lorenz, Carles Bona-Casas, Bastien Chopard, and Alfons G Hoekstra. Foundations of distributed multiscale computing: Formalization, specification, and analysis. Journal of Parallel and Distributed Computing, 73(4):465–483, 2013.
Derek Groen, Elena Vlahu-Gjorgievska, Huiru Zheng, Mihnea Alexandru Moisescu and Ivan Chorbev
83 Phase-Field Based Simulations of Embryonic Branching Morphogenesis [abstract]
Abstract: The mechanism that controls embryonic branching is not fully understood. Of all proposed mechanism, only a Turing pattern-based model succeeds in predicting the location of newly emerging branches during lung and kidney branching morphogenesis. Turing models are based on at least two coupled non-linear reaction-diffusion equations. In case of the lung model, the two components (ligands and receptors) are produced in two different tissue layers [1]. Thus the ligand is produced in the outer mesenchymal layer and the receptor is produced in the inner, branching epithelial layer; the diffusion of receptors is restricted to this epithelial layer. So far, numerical instabilities due to highly complex mesh deformations limit the maximal rounds of branching that can be simulated in an ALE-based framework. A recently developed Phase-Field-based framework [2], shows promising results for the simulation of consecutive 3D branching events. In this talk, I will present our Phase-Field-based framework for simulating the inner epithelial and the outer mesenchymal layer, how we coupled the reaction-diffusion equations to the diffuse / implicit domain as well as the incorporation of additional equations representing further components influencing the growth of a mammalian lung. [1] D. Menshykau et al., "An interplay of geometry and signaling enables robust lung branching morphogenesis.", Development 141(23): 4526-4536, 2014 [2] LD. Wittwer et al., "Simulating Organogenesis in COMSOL: Phase-Field Based Simulations of Embryonic Lung Branching Morphogenesis.", Proceedings of the 2016 COMSOL Conference in Munich, 2016
Lucas D. Wittwer, Sebastian Aland and Dagmar Iber

Workshop on Computational Optimization,Modelling and Simulation (COMS) Session 2

Time and Date: 15:45 - 17:25 on 12th June 2017

Room: HG D 5.2

Chair: Xin-She Yang

572 A Hybrid Heuristic in GPU-CPU Based on Scatter Search for the Generalized Assignment Problem [abstract]
Abstract: In the Generalized Assignment Problem (GAP), tasks must be allocated to machines with limited resources, in order to minimize processing costs. This problem has several industrial applications and often appears as a substructure in other combinatorial optimization problems. We propose a hybrid method inspired by Scatter Search metaheuristic, that efficiently generates a pool of solutions using a Tabu list criteria and an Ejection Chain mechanism. Common characteristics are extracted from the pool and solutions are combined by exploring a restricted search space, as a Binary Programming (BP) model. This method was implemented as a parallel approach to run in a Graphics Processing Unit (GPU). Experimental results show that the proposed method is very competitive to the algorithms found in the literature. On average, a gap of 0.09% is obtained over a set of 45 instances, when compared to lower bounds. Due to the integration of the method with an exact BP solver, it was capable of proving the optimality of small size instances, also finding new best known solutions for 21 instances. In terms of computational times, the proposed method performs on average 8 times faster than literature, also indicating that the proposed approach is scalable and robust for practical applications.
Danilo Santos Souza, Haroldo Gambini Santos and Igor Machado Coelho
314 An Exact Resolution for the Probabilistic Traveling Salesman Problem under the A Priori Strategy [abstract]
Abstract: The Probabilistic Traveling Salesman Problem (PTSP) is an extension of the classical Traveling Salesman Problem (TSP). The main difference is the stochastic presence of the customers, that is, the number of them to be visited each time is a random variable, where each customer associates with a given probability of occurrence. The resolution of problem consists in finding an a priori tour visiting all customers which minimizes the expected length over all possibilities. In this paper, we propose an exact algorithm for the solution of the PTSP. In this context, we derive new expressions for the lower bound and transitional, permanent evaluations as well as the mechanism of separation. The numerical results demonstrate the advantage of the proposed evaluations.
Mohamed Abdellahi Amar, Walid Khaznaji and Monia Bellalouna
97 A Matrix Approach to DC Railway Electrification Verification [abstract]
Abstract: There are some rules for 3,000 V DC electrification in the network ruled by the Spanish Railway Infrastructure Authority (ADIF). As far as we know, the correction of the installations is nowadays manually checked by an expert, as used to be done for expert systems years ago. We propose a computer tool that is an aid for the expert in checking that the positioning of the insulators, circuit breakers, load disconnectors, feeders,... fulfills the requirements established. The computer tool allows the expert to automatically check the sections fed in the different scenarios proposed in the requirements.
Eugenio Roanes-Lozano and Ruben Gonzalez-Martin
284 A Multi-Objective Approach to the Competitive Facility Location Problem [abstract]
Abstract: We introduce a novel approach for a competitive facility location problem in which multiple competitors aim to gain market share. The problem is called the Competitive Maximal Covering Location Problem (CMCLP) based on the classical Maximal Covering Location Problem. Typically, the CMCLP is modeled as a Stackelberg game in which the first player and then the other one locate a fixed number of facilities. On the other hand, the present work considers multiple competitors, and the objective is on discovering a set of the competitors’ decision tuples that are not dominated by any other decision tuples in the solution space. Thereby, our modeling paradigm aims to help competing firms understand tradeoffs when they engage in negotiations. We introduce a mathematical model for the CMCLP with two competitors and a genetic algorithm to solve the problems with multiple competitors. We present computational experiments to demonstrate that the proposed algorithm is able to approximate the true Pareto front.
Abdullah Konak, Sadan Kulturel-Konak and Lawrence Snyder
578 Multi-objective optimisation in scientific workflows [abstract]
Abstract: Engineering design is typically a complex process that involves finding a set of designs satisfying various performance criteria. As a result, optimisation algorithms dealing with only single-objective are not sufficient to deal with many real-life problems. Meanwhile, scientific workflows have been shown to be an effective technology for automating and encapsulating scientific processes. While optimisation algorithms have been integrated into workflow tools, they are generally single-objective. This paper first presents our latest development to incorporate multi-objective optimisation algorithms into scientific workflows. We demonstrate the efficacy of these capabilities with the formulation of a three-objective aerodynamics optimisation problem. We target to improve the aerodynamic characteristics of a typical 2D airfoil profile considering also the laminar-turbulent transition location for more accurate estimation of the total drag. We deploy two different heuristic optimisation algorithms and compare the preliminary results.
Hoang Nguyen, Zane van Iperen, Sreekanth Raghunath, David Abramson, Timoleon Kipouros and Sandeep Somasekharan

Computational Chemistry and its Applications (CCA) Session 1

Time and Date: 15:45 - 17:25 on 12th June 2017

Room: HG F 33.1

Chair: Luthi Hans Peter

591 Quantum chemistry and large scale computations: A connected parallel development [abstract]
Abstract: Beginning with the famous 1929 statement of Paul Dirac concerning the complexity of the application of the fundamental equations of quantum mechanics to chemistry, computers and quantum chemistry have been joined at the hip. Indeed some of the most important developments in computer technology (for example, what was once widely known as 'time sharing”) have been furthered by quantum chemists. This lecture will look at developments in both fields of endeavor from 1940 to the present.
Henry Schaefer
591 Quantum chemistry and large scale computations: A connected parallel development [abstract]
Abstract: Beginning with the famous 1929 statement of Paul Dirac concerning the complexity of the application of the fundamental equations of quantum mechanics to chemistry, computers and quantum chemistry have been joined at the hip. Indeed some of the most important developments in computer technology (for example, what was once widely known as 'time sharing”) have been furthered by quantum chemists. This lecture will look at developments in both fields of endeavor from 1940 to the present.
Henry Schaefer
283 Monte Carlo Study of the Cyrstalline and Amorphous NaK Alloy [abstract]
Abstract: Metropolis Monte Carlo simulations of the eutectic NaK alloy are performed using the Second Moment Approximation (SMA) model potential across a wide range of temperatures at constant pressure. The alloy structure and thermodynamics are analyzed along with the atomic level structures using a variety of structure identification methods. Both enthalpy and density are followed along an annealing process that reveals a clear melting point around 260 K. At lower temperatures, two thermodynamic branches are identified as crystalline and amorphous solids.
Estela Blaisten-Barojas and Doug Reitz
395 Towards a better understanding of on and off target effects of the lymphocyte-specific kinase LCK for the development of novel and safer pharmaceuticals [abstract]
Abstract: In this work we have developed a multi-tiered computational platform to study protein-drug interactions. At the beginning of the workflow more efficient and less accurate methods are used to enable large libraries of proteins in many conformations and massive chemical libraries to be screened. At each subsequent step in the workflow a subset of input data is investigated with increased accuracy and more computationally expensive methods. In this work we demonstrate the developed workflow with the investigation of the lymphocyte-specific kinase LCK, which is implicated as a drug target in many cancers and also known to have toxic effects when unintentionally targeted.
Xiaofei Zhang, Amir Kucharski, Sally R. Ellingson and Wibe A. de Jong
616 Development of a data-centric stability test for iodanes by using the support vector machine [abstract]
Abstract: Hypervalent iodine compounds, in particular trivalent iodine species (lambda-3 iodanes) have become important reagents for the transfer of electrophilic substituents to arenes and other nucleophiles. Quantum chemical studies have shown that iodanes have a rich and often complex reactivity, reminiscent of the one of organometallics. We further explored the reactivity of these compounds based on data-centric models. A large array of iodanes with different electrophilic substituents, different leaving groups and bridges was created and investigated statistically. In this presentation, we show that the kinetic and thermodynamic stability of these reagents can be described by simple structural parameters such as bond lengths. This allows us to develop a stability test for these reagents by using a machine learning technique, e.g., the support vector machine.
Shungo Koichi and Hans Peter Lüthi

Urgent Computing: Computations for Decision Support in Critical Situations (UC) Session 1

Time and Date: 15:45 - 17:25 on 12th June 2017

Room: HG E 33.3

Chair: Alexander Boukhanovsky

521 Simulation of emergency care for patients with ACS in Saint Petersburg for ambulance decision making [abstract]
Abstract: One of the stages of emergency medical care in case of ACS (if there are medical conditions for surgical intervention) is directly linked to the time between the first contact with the patient and the and inflating the balloon in the coronary artery (in a medical institution). Time of the operation start a medical facility depends on the time of patient delivery to hospital, as well as on the waiting time in the queue in the institution. This paper describes a development of ambulance model for obtaining aggregate estimation of these 2 periods of time. The estimation time is obtained by means of described in the article the decision support system (DSS) in the ambulance service. Unlike modern navigation systems DSS takes into account ambulance vehicle behavior (the ability to exit into oncoming traffic) and availability of free operation rooms. With the help of the described simulation model of the ambulance service we carried out the time distribution analysis (between the first contact with the patient and surgical intervention in case of ACS) in St. Petersburg, Russia. Simulation scenario uses real data on the work of ambulance service in the city.
Ivan Derevitskiy, Evgeniy Krotov, Daniil Voloshin, Alexey Yakovlev, Sergey V. Kovalchuk and Vladislav Karbovskii
362 Smart levee monitoring and flood decision support system: reference architecture and urgent computing management [abstract]
Abstract: Real-time disaster management and decision support systems rely on complex deadline-driven simulations and require advanced middleware services to ensure that the requested deadlines are met. In this paper, we propose a reference architecture of an integrated smart levee monitoring and flood decision support system, focusing on the decision support workflow and urgent computing management. The architecture is implemented in project ISMOP in which controlled flooding experiments are conducted using a full-scale experimental smart levee. While the system operating in the ISMOP project monitors a~test levee, it is designed to be scalable to large-scale flood scenarios.
Bartosz Balis, Tomasz Bartyński, Marian Bubak, Daniel Harezlak, Marek Kasztelnik, Maciej Malawski, Piotr Nowakowski, Maciej Pawlik and Bartosz Wilk
565 Firemap: Dynamic Data-Driven Predictive Wildfire Modeling and Visualization Environment [abstract]
Abstract: Wildfires are destructive fires over the wildland that can wipe out large areas of vegetation and infrastructure. Such fires are hard to control and manage as they can change directions almost instantly, driven by changing environmental conditions. Effective response to such events require ability to monitor and predict the behavior of the fire as fast as they change. The WIFIRE project builds an end-to-end cyberinfrastructure for real-time and data-driven simulation, prediction, and visualization of wildfire behavior. One goal of WIFIRE is to provide the tools to predict a more accurate rate of a spreading wildfire. To this end, WIFIRE has developed interfaces for ingesting and visualizing high-density sensor networks to improve fire and weather predictions, and has created a data model for wildfire resources including sensed and archived data, sensors, satellites, cameras, modeling tools, workflows, and social information including Twitter feeds for wildfire research and response. This paper presents WIFIRE's Firemap web platform to make these geospatial data and products accessible. Through a web browser, Firemap enables geospatial information visualization and a unified access to geospatial workflows using Kepler. Using GIS capabilities combined with scalable big data integration and processing, Firemap enables simple execution of the model with options for running ensembles by taking the information uncertainty into account. The results are easily viewable, sharable, repeatable, and can be animated as a time series.
Daniel Crawl, Jessica Block, Kai Lin and Ilkay Altintas
593 Performance-aware scheduling of streaming applications using genetic algorithm [abstract]
Abstract: The main objective of Decision Support Systems is detection of critical states and response on them in time. Such systems can be based on constant monitoring of continuously incoming data. Stream processing is carried out on the basis of computing infrastructure and specialized frameworks such as Apache Storm, Flink, Spark Streaming. However, to provide the necessary system performance at high load incoming data, additional data processing mechanisms are required. In particular, the efficient scheduling of streaming applications plays an important role in the data stream processing. Therefore, this paper is devoted to investigation of genetic algorithm to improve the performance of data stream processing system. The proposed genetic algorithm is developed and integrated into Apache Storm platform, and its efficiency is compared with heuristic algorithm for scheduling of Storm streaming applications.
Pavel Smirnov, Mikhail Melnik and Denis Nasonov
365 Towards an operational database for real-time environmental monitoring and early warning systems [abstract]
Abstract: Real-time environmental monitoring, early warning and decision support systems (EMEWD) require advanced management of operational data, i.e. recent sensor data required for the assessment of the current situation. In this paper, we evaluate the suitability of four data models and corresponding database technologies -- Document database MongoDB, relational database PostgreSQL, Dictionary data server Redis, and time series database InfluxDB -- to serve as an operational database for EMEWD systems. For each of the evaluated databases, we design alternative data models to represent time series data, and experimentally evaluate them. Then we perform comparative performance evaluation of all databases, each using the best model. We designed performance tests reflecting realistic conditions, using mixed workloads (simultaneous read and write operations) and queries typical for smart levee monitoring and flood decision support system. Overall the results of the experiments allow us to answer interesting questions, such as: (1) how to best implement time series in a~given data model? (2) What are the reasonable limits of the operational database sizes? (3) What are the performance limits for different databases?
Bartosz Balis, Marian Bubak, Daniel Harezlak, Piotr Nowakowski, Maciej Pawlik and Bartosz Wilk

Applications of Matrix Computational Methods in the Analysis of Modern Data (AMCMD) Session 2

Time and Date: 15:45 - 17:25 on 12th June 2017

Room: HG E 41

Chair: Kourosh Modarresi

569 Separable Covariance Matrices and Kronecker Approximation [abstract]
Abstract: Abstract: When a model structure allows for the error covariance matrix to be written in the form of the Kronecker product of two positive definite covariance matrices, the estimation of the relevant parameters is intuitive and easy to carry out. In many time series models, the covariance matrix does not have a separable structure. Van Loan and Pitsanis (1993) provide an approximation with Kronecker products.In this paper, we apply their method to estimate the parameters of a multivariate regression model withautoregressive errors. An illustrative example is also provided
Raja Velu and Krzysztof Herman
137 Distributed Bayesian Probabilistic Matrix Factorization [abstract]
Abstract: Using the matrix factorization technique in machine learning is very common mainly in areas like recommender systems. Despite its high prediction accuracy and its ability to avoid over-fitting of the data, the Bayesian Probabilistic Matrix Factorization algorithm (BPMF) has not been widely used on large scale data because of the prohibitive cost. In this paper, we propose a distributed high-performance parallel implementation of the BPMF using Gibbs sampling on shared and distributed architectures. We show by using efficient load balancing using work stealing on a single node, and by using asynchronous communication in the distributed version we beat state of the art implementations.
Tom Vander Aa, Imen Chakroun and Tom Haber
551 Finding the Winner of a Hypothesis Test via Sample Allocation Optimization [abstract]
Abstract: Hypothesis testing, which is one of the most important and practical tools for comparing different distributions, has been recently drawn a huge interest among the practitioners. In many real-life scenarios, among many different options, one wants to find the best option in the quickest time. Statistical hypothesis tests were previous approaches for addressing this problem and had shown to achieve a satisfactory performance in the presence of large number of samples. Multi-armed bandit methods which were first introduced in the mid-twenty century, has been recently shown promising performances for this task as well. Although bandit methods work well for regret minimization, they are not well-designed for finding the winner of a campaign with some statistical guarantee. In this paper, we will introduce a new algorithm based on the sample allocation optimization, which finds the winner of the campaign with statistical guarantees, in the quickest time.
Kourosh Modarresi and Khashayar Khosravi
575 Efficient Hybrid Algorithms for Computing Clusters Overlap [abstract]
Abstract: Every year, marketers target different segments of the population with multitudes of advertisements. However, millions of dollars are wasted targeting similar segments with different advertisements. Furthermore, it is extremely expensive to compute the similarity between the segments because these segments can be very large, on the order of millions and billions. In this project, we aim to come up with a fast algorithm that can determine the similarity between large segments with a higher degree of accuracy than other known methods.
Kourosh Modarresi, Aran Nayebi and Yi Liu
520 Optimizing Fiber to the Premises Build-Out Plans Using Geoqueries and Spectral Graph Paritioning [abstract]
Abstract: The build-out of fiber internet connectivity to the premises (FTTP) is a large scale, capital intensive endeavor. We developed a method that allows us to take advantage of economies of scale by finding regions of adjacent Distribution Areas with high similarity in cost structure. The algorithm employs geographic queries and spectral clustering on weighted large scale adjacency graphs to find these high adjacency regions. We present the business context, development of a similarity measure based on build out cost structure, the application of spectral clustering on the graphs, and post processing steps.
Susanne Halstead, Halim Damerdji and Andrew Domenico