Large Scale Scientific Computing 11th

International Conference LSSC 2017

Sozopol Bulgaria June 5 9 2017 Revised
Selected Papers 1st Edition Ivan Lirkov
Visit to download the full and correct content document:
https://textbookfull.com/product/large-scale-scientific-computing-11th-international-co
nference-lssc-2017-sozopol-bulgaria-june-5-9-2017-revised-selected-papers-1st-editi
on-ivan-lirkov/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Algorithmic Aspects of Cloud Computing Third

International Workshop ALGOCLOUD 2017 Vienna Austria
September 5 2017 Revised Selected Papers 1st Edition
Dan Alistarh
https://textbookfull.com/product/algorithmic-aspects-of-cloud-
computing-third-international-workshop-algocloud-2017-vienna-
austria-september-5-2017-revised-selected-papers-1st-edition-dan-
alistarh/

Information Security and Cryptology: 13th International

Conference, Inscrypt 2017, Xi'an, China, November 3–5,
2017, Revised Selected Papers 1st Edition Xiaofeng Chen

https://textbookfull.com/product/information-security-and-
cryptology-13th-international-conference-inscrypt-2017-xian-
china-november-3-5-2017-revised-selected-papers-1st-edition-
xiaofeng-chen/

Cloud Computing and Service Science 7th International

Conference CLOSER 2017 Porto Portugal April 24 26 2017
Revised Selected Papers Donald Ferguson

https://textbookfull.com/product/cloud-computing-and-service-
science-7th-international-conference-closer-2017-porto-portugal-
april-24-26-2017-revised-selected-papers-donald-ferguson/

Internet Multimedia Computing and Service 9th

International Conference ICIMCS 2017 Qingdao China
August 23 25 2017 Revised Selected Papers 1st Edition
Benoit Huet
https://textbookfull.com/product/internet-multimedia-computing-
and-service-9th-international-conference-icimcs-2017-qingdao-
china-august-23-25-2017-revised-selected-papers-1st-edition-
Studies on Speech Production 11th International Seminar
ISSP 2017 Tianjin China October 16 19 2017 Revised
Selected Papers Qiang Fang

https://textbookfull.com/product/studies-on-speech-
production-11th-international-seminar-issp-2017-tianjin-china-
october-16-19-2017-revised-selected-papers-qiang-fang/

Digital TV and Wireless Multimedia Communication: 14th

International Forum, IFTC 2017, Shanghai, China,
November 8-9, 2017, Revised Selected Papers 1st Edition
Guangtao Zhai
https://textbookfull.com/product/digital-tv-and-wireless-
multimedia-communication-14th-international-forum-
iftc-2017-shanghai-china-november-8-9-2017-revised-selected-
papers-1st-edition-guangtao-zhai/

Computational Aspects and Applications in Large Scale

Networks NET 2017 Nizhny Novgorod Russia June 2017
Valery A. Kalyagin

https://textbookfull.com/product/computational-aspects-and-
applications-in-large-scale-networks-net-2017-nizhny-novgorod-
russia-june-2017-valery-a-kalyagin/

Critical Information Infrastructures Security: 12th

International Conference, CRITIS 2017, Lucca, Italy,
October 8-13, 2017, Revised Selected Papers Gregorio
D'Agostino
https://textbookfull.com/product/critical-information-
infrastructures-security-12th-international-conference-
critis-2017-lucca-italy-october-8-13-2017-revised-selected-
papers-gregorio-dagostino/

Fundamentals of Software Engineering 7th International

Conference FSEN 2017 Tehran Iran April 26 28 2017
Revised Selected Papers 1st Edition Mehdi Dastani

https://textbookfull.com/product/fundamentals-of-software-
engineering-7th-international-conference-fsen-2017-tehran-iran-
april-26-28-2017-revised-selected-papers-1st-edition-mehdi-
 Ivan Lirkov
Svetozar Margenov (Eds.)
LNCS 10665

Large-Scale
Scientific Computing
11th International Conference, LSSC 2017
Sozopol, Bulgaria, June 5–9, 2017
Revised Selected Papers

123
Lecture Notes in Computer Science 10665
Commenced Publication in 1973
Founding and Former Series Editors:
Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen

Editorial Board
David Hutchison
Lancaster University, Lancaster, UK
Takeo Kanade
Carnegie Mellon University, Pittsburgh, PA, USA
Josef Kittler
University of Surrey, Guildford, UK
Jon M. Kleinberg
Cornell University, Ithaca, NY, USA
Friedemann Mattern
ETH Zurich, Zurich, Switzerland
John C. Mitchell
Stanford University, Stanford, CA, USA
Moni Naor
Weizmann Institute of Science, Rehovot, Israel
C. Pandu Rangan
Indian Institute of Technology, Madras, India
Bernhard Steffen
TU Dortmund University, Dortmund, Germany
Demetri Terzopoulos
University of California, Los Angeles, CA, USA
Doug Tygar
University of California, Berkeley, CA, USA
Gerhard Weikum
Max Planck Institute for Informatics, Saarbrücken, Germany
More information about this series at http://www.springer.com/series/7407
Ivan Lirkov Svetozar Margenov (Eds.)
•

Large-Scale
Scientific Computing
11th International Conference, LSSC 2017
Sozopol, Bulgaria, June 5–9, 2017
Revised Selected Papers

123
Editors
Ivan Lirkov Svetozar Margenov
Institute of Information and Communication Institute of Information and Communication
Technologies Technologies
Bulgarian Academy of Sciences Bulgarian Academy of Sciences
Sofia Sofia
Bulgaria Bulgaria

ISSN 0302-9743 ISSN 1611-3349 (electronic)

Lecture Notes in Computer Science
ISBN 978-3-319-73440-8 ISBN 978-3-319-73441-5 (eBook)
https://doi.org/10.1007/978-3-319-73441-5

Library of Congress Control Number: 2017962887

LNCS Sublibrary: SL1 – Theoretical Computer Science and General Issues

© Springer International Publishing AG 2018

Chapters 15 and 16 was created within the capacity of an US governmental employment. US copyright
protection does not apply.
The chapter “Parallel Aggregation Based on Compatible Weighted Matching for AMG” is licensed under the
terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/
by/4.0/). For further details see license information in the chapter.
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now
known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are
believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors
give a warranty, express or implied, with respect to the material contained herein or for any errors or
omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
 Preface

The 11th International Conference on Large-Scale Scientific Computations (LSSC

2017) was held in Sozopol, Bulgaria, June 5–9, 2017. The conference was organized
by the Institute of Information and Communication Technologies at the Bulgarian
Academy of Sciences in cooperation with Society for Industrial and Applied Mathe-
matics (SIAM) and Sozopol municipality.
The plenary invited speakers and lectures were:
– J. Adler, “Energy-Minimization, Finite Elements, and Multilevel Methods for
Liquid Crystals”
– J. Gopalakrishnan, “Theoretical and Practical Aspects of DPG Methods”
– U. Langer, “Multi-Patch Discontinuous Galerkin Space and Space-Time Isogeo-
metric Analysis”
– J. Pasciak, “Numerical Approximation of a Space-Time Fractional Parabolic
Equation”
– G. G. Yin, “Numerical Methods for Consensus of Networked Systems and
Stochastic Control with Soft Constraints”
The success of the conference and the present volume are the outcome of the joint
efforts of many partners from various institutions and organizations. First, we would
like to thank all the members of the Scientific Committee for their valuable contribution
forming the scientific face of the conference, as well as for their help in reviewing
contributed papers. We especially thank the organizers of the special sessions. We are
also grateful to the staff involved in the local organization.
Traditionally, the purpose of the conference is to bring together scientists working
with large-scale computational models in natural sciences and environmental and
industrial applications, and specialists in the field of numerical methods and algorithms
for modern high-performance computers. The invited lectures reviewed some of the
most advanced achievements in the field of numerical methods and their efficient
applications. The conference talks were presented by researchers from academic
institutions and practical industry engineers including applied mathematicians,
numerical analysts, and computer experts. The general theme for LSSC 2017 was
“Large-Scale Scientific Computing” with a particular focus on the organized special
sessions.
The special sessions and organizers were:
– Space-Time Methods for Solving Time-Dependent PDEs — U. Langer
– Advanced Discretizations and Solvers for Coupled Systems of Partial Differential
Equations — J. Adler, X. Hu, L. Zikatanov
– Least-Squares Finite Element Methods — F. Bertrand and P. Bochev
– Advances in Heterogeneous Numerical Methods for Multi Physics Problems
— P. Bochev, M. Perego
VI Preface

– Advanced Numerical Methods for Nonlinear Elliptic Partial Differential Equations

— J. Kraus
– Solvers and Error Estimators for Mixed Finite Elements in Solid Mechanics
— G. Starke
– Control and Optimization of Dynamical Systems — M. Krastanov, V. Veliov
– HPC and Big Data: Algorithms and Applications — A. Karaivanova, T. Gurov,
E. Atanassov
– Toward Exascale Computation — O. Iliev
– Monte Carlo Methods: Theory, Applications and Distributed Computing
— I. Dimov, R. Georgieva, M. Nedjalkov
– Application of Metaheuristics to Large-Scale Problems — S. Fidanova, G. Luque
– Large-Scale Models: Numerical Methods, Parallel Computations and Applications
— K. Georgiev, Z. Zlatev
– Large-Scale Numerical Computations for Sustainable Energy Production and
Storage — P. D’Ambra, D. di Serafino, S. Filippone
About 150 participants from all over the world attended the conference representing
some of the strongest research groups in the field of advanced large-scale scientific
computing. This volume contains 66 papers by authors from 21 countries.
The next international conference LSSC will be organized in June 2019.

November 2017 Ivan Lirkov

Svetozar Margenov
 Organization

Scientific Committee
James Adler Tufts University, USA
Pasqua D’Ambra Istituto per le Applicazioni del Calcolo Mauro Picone,
CNR, Naples, Italy
Emanouil Atanassov Institute of Information and Communication Technologies,
BAS, Bulgaria
Fleurianne Bertrand Universität Duisburg-Essen, Germany
Pavel Bochev Sandia National Laboratories, USA
Ivan Dimov Institute of Information and Communication Technologies,
BAS, Bulgaria
Stefka Dimova Sofia University, Bulgaria
Stefka Fidanova Institute of Information and Communication Technologies,
BAS, Bulgaria
Salvatore Filippone Cranfield University, UK
Krassimir Georgiev Institute of Information and Communication Technologies,
BAS, Bulgaria
Jay Gopalakrishnan Portland State University, USA
Todor Gurov Institute of Information and Communication Technologies,
BAS, Bulgaria
Xiaozhe Hu Tufts University, USA
Oleg Iliev ITWM, Germany
Aneta Karaivanova Institute of Information and Communication Technologies,
BAS, Bulgaria
Mikhail Krastanov Sofia University, Bulgaria
Johannes Kraus University of Duisburg-Essen, Germany
Ulrich Langer Johannes Kepler University Linz, Austria
Raytcho Lazarov Texas A&M University, USA
Ivan Lirkov Institute of Information and Communication Technologies,
BAS, Bulgaria
Gabriel Luque University of Málaga, Spain
Svetozar Margenov Institute of Information and Communication Technologies,
BAS, Bulgaria
Joseph Pasciak Texas A&M University, College Station, USA
Mauro Perego Sandia National Laboratories, USA
Daniela di Serafino Università degli Studi della Campania Luigi Vanvitelli,
Italy
Gerhard Starke University of Duisburg-Essen, Germany
Vladimir Veliov TU Vienna, Austria
VIII Organization

Gang George Yin Wayne State University, USA

Zahari Zlatev Aarhus University, Denmark
Ludmil Zikatanov Pennsylvania State University, USA
 Contents

Invited Papers

Discrete Energy Laws for the First-Order System Least-Squares

Finite-Element Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
J. H. Adler, I. Lashuk, S. P. MacLachlan, and L. T. Zikatanov

Multipatch Space-Time Isogeometric Analysis of Parabolic

Diffusion Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
U. Langer, M. Neumüller, and I. Toulopoulos

Numerical Methods for Controlled Switching Diffusions . . . . . . . . . . . . . . . 33

G. Yin, C. Zhang, and L. Y. Wang

Space-Time Methods for Solving Time-Dependent PDEs

Preconditioners for Time-Harmonic Optimal Control

Eddy-Current Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Owe Axelsson and Dalibor Lukáš

Functional Type Error Control for Stabilised Space-Time IgA

Approximations to Parabolic Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Ulrich Langer, Svetlana Matculevich, and Sergey Repin

An Algebraic Multigrid Method for an Adaptive Space–Time Finite

Element Discretization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
Olaf Steinbach and Huidong Yang

Advanced Discretizations and Solvers for Coupled Systems

of Partial Differential Equations

Splitting Schemes for Mixtures of Nematic-Isotropic Flows

with Anchoring Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Giordano Tierra, Francisco Guillén-González,
and María Ángeles Rodríguez-Bellido

Two Classes of Vector Domain Decomposition Schemes

for Time-Dependent Problems with Overlapping Subdomains . . . . . . . . . . . . 85
Petr N. Vabishchevich
X Contents

Least-Squares Finite Element Methods

An Alternative Proof of a Strip Estimate for First-Order System

Least-Squares for Interface Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Fleurianne Bertrand

Spectral Mimetic Least-Squares Method for Div-curl Systems. . . . . . . . . . . . 103

Marc Gerritsma and Artur Palha

Spectral Mimetic Least-Squares Methods on Curvilinear Grids . . . . . . . . . . . 111

R. O. Hjort and B. Gervang

Spectral Mimetic Least-Squares Method for Curl-curl Systems . . . . . . . . . . . 119

Artur Palha and Marc Gerritsma

Numerical Solution of Cahn-Hilliard System by Adaptive Least-Squares

Spectral Element Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
Keunsoo Park, Marc Gerritsma, and Maria Fernandino

Stress-Velocity Mixed Least-Squares FEMs for the Time-Dependent

Incompressible Navier-Stokes Equations . . . . . . . . . . . . . . . . . . . . . . . . . . 137
Alexander Schwarz, Carina Nisters, Solveigh Averweg,
and Jörg Schröder

Advances in Heterogeneous Numerical Methods for Multi

Physics Problems

A Virtual Control Coupling Approach for Problems with Non-coincident

Discrete Interfaces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Pavel Bochev, Paul Kuberry, and Kara Peterson

Towards a Scalable Multifidelity Simulation Approach for Electrokinetic

Problems at the Mesoscale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
Brian D. Hong, Mauro Perego, Pavel Bochev, Amalie L. Frischknecht,
and Edward G. Phillips

Advanced Numerical Methods for Nonlinear Elliptic Partial

Differential Equations

On a Problem of Optimal Control of Convection-Diffusion Processes . . . . . . 167

Aigul Manapova and Fedor Lubyshev

Verifications of Primal Energy Identities for Variational Problems

with Obstacles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Sergey Repin and Jan Valdman
 Contents XI

Control and Optimization of Dynamical Systems

An Optimal Control Problem with a Risk Zone . . . . . . . . . . . . . . . . . . . . . 185

Sergey M. Aseev

Spreading Rumors and External Actions . . . . . . . . . . . . . . . . . . . . . . . . . . 193

Séverine Bernard, Ténissia César, and Alain Piétrus

Superposition Principle for Differential Inclusions . . . . . . . . . . . . . . . . . . . . 201

Giulia Cavagnari, Antonio Marigonda, and Benedetto Piccoli

Estimation of Star-Shaped Reachable Sets of Nonlinear Control Systems . . . . 210

Tatiana F. Filippova

On Reachability Analysis of Nonlinear Systems with Joint

Integral Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
Mikhail Gusev

Existence Theorem for Infinite Horizon Optimal Control Problems

with Mixed Control-State Isoperimetrical Constraint . . . . . . . . . . . . . . . . . . 228
Valeriya Lykina

On the Regularity of Linear-Quadratic Optimal Control Problems

with Bang-Bang Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
J. Preininger, T. Scarinci, and V. M. Veliov

HPC and Big Data: Algorithms and Applications

On Monte Carlo and Quasi-Monte Carlo for Matrix Computations . . . . . . . . 249

Vassil Alexandrov, Diego Davila, Oscar Esquivel-Flores,
Aneta Karaivanova, Todor Gurov, and Emanouil Atanassov

On the Parallel Implementation of Quasi-Monte Carlo Algorithms. . . . . . . . . 258

E. Atanassov, T. Gurov, S. Ivanovska, A. Karaivanova,
and T. Simchev

TVRegCM Numerical Simulations - Preliminary Results . . . . . . . . . . . . . . . 266

Georgi Gadzhev, Vladimir Ivanov, Kostadin Ganev,
and Hristo Chervenkov

Territorial Design Optimization for Business Sales Plan . . . . . . . . . . . . . . . . 275

Laura Hervert-Escobar and Vassil Alexandrov

Monte Carlo Algorithms for Problems with Partially

Reflecting Boundaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
Nikolai A. Simonov
XII Contents

Toward Exascale Computation

Renormalization Based MLMC Method for Scalar Elliptic SPDE . . . . . . . . . 295

Oleg Iliev, Jan Mohring, and Nikolay Shegunov

Performance Analysis of MG Preconditioning on Intel Xeon Phi:

Towards Scalability for Extreme Scale Problems with
Fractional Laplacians . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
Nikola Kosturski, Svetozar Margenov, and Yavor Vutov

Application of Metaheuristics to Large-Scale Problems

Training Feed-Forward Neural Networks Employing Improved

Bat Algorithm for Digital Image Compression . . . . . . . . . . . . . . . . . . . . . . 315
Adis Alihodzic

Modeling and Optimization of Pickup and Delivery Problem

Using Constraint Logic Programming . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
Amelia Bădică, Costin Bădică, Florin Leon, and Ion Buligiu

Intercriteria Analysis over Intuitionistic Fuzzy Data . . . . . . . . . . . . . . . . . . . 333

Veselina Bureva, Evdokia Sotirova, Vassia Atanassova,
Nora Angelova, and Krassimir Atanassov

Genetic Algorithm with Optimal Recombination for the Asymmetric

Travelling Salesman Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
Anton V. Eremeev and Yulia V. Kovalenko

Heuristic Algorithm for 2D Cutting Stock Problem . . . . . . . . . . . . . . . . . . . 350

Georgi Evtimov and Stefka Fidanova

Influence of Ant Colony Optimization Parameters

on the Algorithm Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358
Stefka Fidanova and Olympia Roeva

2D Optimal Packing with Population Based Algorithms. . . . . . . . . . . . . . . . 366

Desislava Koleva, Maria Barova, and Petar Tomov

A Non-dominated Sorting Approach to Bi-objective Optimisation

of Mixed-Model Two-Sided Assembly Lines . . . . . . . . . . . . . . . . . . . . . . . 374
Ibrahim Kucukkoc

Development of Threshold Algorithms for a Location Problem

with Elastic Demand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
Tatyana Levanova and Alexander Gnusarev
 Contents XIII

Investigation of Genetic Algorithm Performance Based on Different

Algorithms for InterCriteria Relations Calculation . . . . . . . . . . . . . . . . . . . . 390
Tania Pencheva, Olympia Roeva, and Maria Angelova

Free Search in Multidimensional Space M . . . . . . . . . . . . . . . . . . . . . . . . . 399

Kalin Penev

Generalized Net Model of Adhesive Capsulitis Diagnosing . . . . . . . . . . . . . 408

Simeon Ribagin, Evdokia Sotirova, and Tania Pencheva

Adaptive Multi-agent System Based on Wasp-Like Behaviour

for the Virtual Learning Game Sotirios . . . . . . . . . . . . . . . . . . . . . . . . . . . 416
Dana Simian and Florentin Bota

Hybrid Approach Based on Combination of Backpropagation

and Evolutionary Algorithms for Artificial Neural Networks Training
by Using Mobile Devices in Distributed Computing Environment . . . . . . . . . 425
Iliyan Zankinski, Maria Barova, and Petar Tomov

Large-Scale Models: Numerical Methods, Parallel Computations

and Applications

Solution of the 3D Neutron Diffusion Benchmark by FEM . . . . . . . . . . . . . 435

A. V. Avvakumov, P. N. Vabishchevich, A. O. Vasilev,
and V. F. Strizhov

Precipitation Pattern Estimation with the Standardized Precipitation

Index in Projected Future Climate over Bulgaria . . . . . . . . . . . . . . . . . . . . . 443
Hristo Chervenkov and Valery Spiridonov

Time Discretization/Linearization Approach Based on HOC Difference

Schemes for Semilinear Parabolic Systems of Atmosphere Modelling . . . . . . 450
I. Dimov, J. Kandilarov, V. Todorov, and L. Vulkov

Landslide Hazard, Environmental Dependencies

and Computer Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 458
Nina Dobrinkova and Pierluigi Maponi

On the Winter Wave Climate of the Western Black Sea: The Changes
During the Last 115 Years . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 466
Vasko Galabov and Hristo Chervenkov

Computer Simulations of Atmospheric Composition in Urban Areas.

Some Results for the City of Sofia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 474
Ivelina Georgieva, Georgi Gadzhev, Kostadin Ganev,
and Nikolay Miloshev
XIV Contents

Numerical Simulation of Deformations of Softwood Sawn Timber . . . . . . . . 483

Vladimir N. Glukhikh, Anna Yu. Okhlopkova, and Petr V. Sivtsev

Large Scale Computations in Fluid Dynamics. . . . . . . . . . . . . . . . . . . . . . . 491

Valentin A. Gushchin

A Domain Decomposition Multilevel Preconditioner for Interpolation

with Radial Basis Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
Gundolf Haase, Dirk Martin, Patrick Schiffmann,
and Günter Offner

Sampling in In Silico Biomolecular Studies: Single-Stage Experiments

vs Multiscale Approaches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 507
Nevena Ilieva, Jiaojiao Liu, Xubiao Peng, Jianfeng He, Antti Niemi,
Peicho Petkov, and Leandar Litov

Cultural Heritage RC Structures Environmentally Degradated: Optimal

Seismic Upgrading by Tention-Ties Under Shear Effects . . . . . . . . . . . . . . . 516
A. Liolios, K. Liolios, A. Moropoulou, K. Georgiev, and I. Georgiev

New Approach to Identifying Solitary Wave Solutions of Modified

Kawahara Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 527
Tchavdar T. Marinov and Rossitza S. Marinova

Sequential Variational Data Assimilation Algorithms at the Splitting

Stages of a Numerical Atmospheric Chemistry Model . . . . . . . . . . . . . . . . . 536
Alexey Penenko, Vladimir Penenko, Elena Tsvetova, Anastasia Grishina,
and Pavel Antokhin

Computational Modelling of the Full Length hIFN-c Homodimer . . . . . . . . . 544

Peicho Petkov, Elena Lilkova, Nevena Ilieva, Genoveva Nacheva,
Ivan Ivanov, and Leandar Litov

Using Advanced Mathematical Tools in Complex Studies Related

to Climate Changes and High Pollution Levels . . . . . . . . . . . . . . . . . . . . . . 552
Zahari Zlatev, Ivan Dimov, Krassimir Georgiev,
and Radim Blaheta

Large-Scale Numerical Computations for Sustainable Energy

Production and Storage

Parallel Aggregation Based on Compatible Weighted Matching

for AMG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 563
Ambra Abdullahi, Pasqua D’Ambra, Daniela di Serafino,
and Salvatore Filippone

Efficient Solution Techniques for Multi-phase Flow in Porous Media . . . . . . 572

Henrik Büsing
 Contents XV

Contributed Papers

A Unified Numerical Approach for a Large Class of Nonlinear

Black-Scholes Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 583
Miglena N. Koleva and Lubin G. Vulkov

Beta-Function B-splines and Subdivision Schemes, a Preliminary Study. . . . . 592

Arne Lakså

Conjugate Gradient Method for Identification of a Spacewise

Heat Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 600
V. I. Vasil’ev, V. V. Popov, and A. M. Kardashevsky

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 609

Invited Papers
 Discrete Energy Laws for the First-Order
System Least-Squares Finite-Element Approach

J. H. Adler1(B) , I. Lashuk1 , S. P. MacLachlan2 , and L. T. Zikatanov3

1
Department of Mathematics, Tufts University, Medford, MA 02155, USA
james.adler@tufts.edu, ilya.lashuk@gmail.com
2
Department of Mathematics and Statistics, Memorial University of Newfoundland,
St. John’s, NL A1C 5S7, Canada
smaclachlan@mun.ca
3
Department of Mathematics, Penn State, University Park, PA 16802, USA
ludmil@psu.edu

Abstract. This paper analyzes the discrete energy laws associated

with first-order system least-squares (FOSLS) discretizations of time-
dependent partial differential equations. Using the heat equation and the
time-dependent Stokes’ equation as examples, we discuss how accurately
a FOSLS finite-element formulation adheres to the underlying energy law
associated with the physical system. Using regularity arguments involv-
ing the initial condition of the system, we are able to give bounds on
the convergence of the discrete energy law to its expected value (zero
in the examples presented here). Numerical experiments are  performed,

showing that the discrete energy laws hold with order O h2p , where
h is the mesh spacing and p is the order of the finite-element space.
Thus, the energy law conformance is held with a higher order than
the expected, O (hp ), convergence of the finite-element approximation.
Finally, we introduce an abstract framework for analyzing the energy
laws of general FOSLS discretizations.

1 Introduction

First-order system least squares (FOSLS) is a finite-element methodology that

aims to reformulate a set of partial differential equations (PDEs) as a system
of first-order equations [8,9]. The problem is posed as a minimization of a func-
tional in which the first-order differential terms appear quadratically, so that
the functional norm is equivalent to a norm meaningful for the given prob-
lem. In equations of elliptic type, this is usually a product H 1 norm. Some of
the compelling features of the FOSLS methodology include: self-adjoint discrete
equations stemming from the minimization principle; good operator conditioning
stemming from the use of first-order formulations of the PDE; and finite-element
and multigrid performance that is optimal and uniform in certain parameters
(e.g., Reynolds number for the Navier-Stokes equations), stemming from uniform
product-norm equivalence.

c Springer International Publishing AG 2018

I. Lirkov and S. Margenov (Eds.): LSSC 2017, LNCS 10665, pp. 3–20, 2018.
https://doi.org/10.1007/978-3-319-73441-5_1
4 J. H. Adler et al.

Successful FOSLS formulations have been developed for a variety of appli-

cations [5,7]. One example of a large-scale physical application is in magneto-
hydrodynamics (MHD) [1–3]. These numerical methods have led to substantial
improvements in MHD simulation technology; however, several important esti-
mates remain to be analyzed to confirm their quantitive accuracy. One of these
is the energy of the system. Using an energetic-variational approach [11–13,15],
energy laws of the MHD system can be derived that show that the total energy
should decay as a direct result of the dissipation in the system. Initial compu-
tations show that the FOSLS method indeed captures this energy law, but it
remains to be shown why it should.
In this paper, we describe the discrete energy laws associated with FOSLS
discretizations of time-dependent PDEs, such as the heat equation or Stokes’
equation, and show quantitatively how they are related to the continuous physi-
cal law. While we only show results for these “simple” linear systems, the results
appear generalizable to more complicated systems, such as MHD. Getting the
correct energy law is not only important for numerical stability, but it is crucial
for capturing the correct physics, especially if singularities or high contrasts in
the solution are present.
The paper is outlined as follows. In Sect. 2, we discuss the energy laws of
a given system and describe their discrete analogues. Section 3 analyzes the
energy laws associated with the FOSLS discretizations of the heat equation,
and the same is done for Stokes’ equations in Sect. 4. For both examples, we
present numerical simulations in Sect. 5. Finally, we discuss a generalization of
the concepts presented here in Sect. 6, and some concluding remarks in Sect. 7.

2 Energy Laws

The energetic-variational approach (EVA) [11–13,15] of hydrodynamic systems

in complex fluids is based on the second law of thermodynamics and relies on
the fundamental principle that the change in the total energy of a system over
time must equal the total dissipation of the system. This energy principle plays
a crucial role in understanding the interactions and coupling between different
scales or phases in a complex fluid. In general, any set of equations that describe
the system can be derived from the underlying energy laws. The energetic vari-
ational principle is based on the energy dissipation law for the whole coupled
system:
∂Etotal
= −D, (1)
∂t
where Etotal is the total energy of the system, and D is the dissipation.
Simple fluids, where we assume no internal (or elastic) energies, can also be
described in this setting and yield the following energy law:
   
∂ 1
|u|2 dx = − ν|∇u|2 dx, (2)
∂t 2 Ω Ω
 Energy Laws for FOSLS 5

where u represents the fluid velocity and ν is the fluid viscosity, accounting for
the dissipation in the system. Applying the so-called least-action principle results
in the integral equation,
 
∂u
+ ∇p, y = ∇ · ν∇u, y , ∀y ∈ V,
∂t
where we assume an incompressible fluid, ∇ · u = 0, and an appropriate Hilbert
space, V. Here, we use ·, · to denote the L2 (Ω) inner product. In strong form,
we obtain the time-dependent Stokes’ equations (assuming appropriate boundary
conditions):
∂u
+ ∇p − ∇ · ν∇u = 0, (3)
∂t
∇ · u = 0. (4)
Note that the energy law can also be derived directly from the PDE itself. First,
we consider the weak form of (3)–(4), multiplying (3) by u and (4) by p and
integrate over Ω. After integration by parts we obtain the following relations:
 
∂u
0= + ∇p − ∇ · ν∇u, u + ∇ · u, p
∂t
 
∂u
= , u + ∇p, u − ∇ · ν∇u, u + ∇ · u, p
∂t
1 ∂
= u, u + ∇p, u + ν∇u, ∇u − u, ∇p .
2 ∂t
Here, we have assumed that the boundary conditions are such that the boundary
terms, resulting from the integration by parts, vanish. Hence, we have
1 ∂
u, u = − ν∇u, ∇u .
2 ∂t
This approach can also be applied to other PDEs, such as the heat equation,
to show similar energy dissipation relations. Let ν be the thermal diffusivity of
the body Ω, and u its temperature. Then the PDE describing the temperature
distribution in Ω is as follows,
∂u
− ∇ · ν∇u = 0, on Ω, u = 0, on ∂Ω. (5)
∂t
As before, we multiply (5) by u and integrate over Ω to obtain that
   
∂u ∂u
0= − ∇ · ν∇u, u = , u − ∇ · ν∇u, u
∂t ∂t
1 ∂
= u, u + ν∇u, ∇u
2 ∂t
Hence,
1 ∂
u, u = − ν∇u, ∇u ,
2 ∂t
which is the scalar version of (2).
6 J. H. Adler et al.

For the remainder of the paper, we analyze (2), specifically how closely the
FOSLS method can approximate the energy law discretely. We will consider
both the scalar (heat equation) and the vector version (Stokes’ equation) in the
numerical results, as the form of the energy law is identical. First, we discuss
how moving to a finite-dimensional space affects the energy law.

3 Heat Equation
First, we consider the heat equation, assuming a constant diffusion coefficient
ν = 1 for simplicity, homogeneous Dirichlet boundary conditions, and a given
initial condition:
∂u(x, t)
= Δu(x, t) ∀x ∈ Ω, ∀t > 0 (6)
∂t
u(x, t) = 0 ∀x ∈ ∂Ω, ∀t ≥ 0 (7)
u(x, 0) = u0 (x) ∀x ∈ Ω̄. (8)
To discretize the problem in time, we consider a symplectic, or energy-conserving,
time-stepping scheme such as Crank-Nicolson. Given a time step size, τ , and time
tn = τ n, we approximate un = u(x, tn ) with the following semi-discrete version
of (6),
un+1 − un Δun+1 + Δun
=
τ 2
To simplify the calculations later, we introduce an intermediate approxima-
tion, un+ 12 , and re-write the semi-discrete problem as
un+ 12 − un
τ  = Δun+ 12
2
un+ 12 (x) = 0 ∀x ∈ ∂Ω, n = 0, 1, 2, . . . (9)
un+1 = 2un+ 12 − un

Remark 1. To obtain the semi-discrete energy law for (9), we perform a similar
procedure as done in Sect. 2, where we multiply the first equation in (9) by
un+ 12 and integrate over the domain. After some simple calculations, we obtain
the corresponding energy law, using L2 −norm notation:
||un+1 ||2 − ||un ||2
= −||∇un+ 12 ||2 (10)
2τ
To use the FOSLS method, we now put the operator into a first-order system.
Since we have reduced the problem to a reaction-diffusion type problem, we
introduce a new vector V = ∇u, and use the H 1 -elliptic equivalent system [8,9]:
⎛ ⎞ ⎛2 ⎞
  −∇ · Vn+ 12 + τ2 un+ 12
un+ 12 τ un
Lτ = ⎝ Vn+ 12 − ∇un+ 12 ⎠ = ⎝ 0 ⎠ . (11)
Vn+ 12
∇ × Vn+ 1 0
2
 Energy Laws for FOSLS 7

Note that Dirichlet boundary conditions on the continuous solution, u, gives rise
to tangential boundary conditions on V, V × n = 0, where n is the normal
vector to the boundary.
Next, we consider a finite-dimensional subspace of a product H 1 space, V h ,
and perform the FOSLS minimization of (11) over V h :
 ⎛ 2 h ⎞
    
 u τ un 
h h
un+ 1 , Vn+ 1 = arg min Lτ − ⎝ 0 
⎠
V ,
2 2
(u,V )∈V h  0 
uhn+1 = 2uhn+ 1 − uhn .
2

For each n, the above minimization results in the following weak set of equations:
  h  ⎛ 2 h⎞ 
un+ 1 τ un
Lτ h
2
− ⎝ 0 ⎠ , Lτ φh = 0 ∀φh ∈ V h , (12)
Vn+ 1
2 0

where the inner products and norms are all in L2 (scalar or vector, depending
on context), unless otherwise noted.
Note, that with the introduction of V, the discrete form of the FOSLS energy
law can now be written,

||uhn+1 ||2 − ||uhn ||2 h

− ||Vn+ 1 || → 0,
2
as h → 0. (13)
2τ 2

The goal of the remainder of this Section is to show how well this energy law is
satisfied. To do so, we make use of the following assumption.
Assumption 1. Assume that the initial condition is smooth enough and the
projection onto the finite-element space has the following property,
 
u0 − uh0  1 ≤ Chp u0 p+1 ,
H H

where p is the order of the finite-element space being considered.

Then, using standard regularity estimates we obtain the following Lemma.
Lemma 1. Let {ui }i=0,1,... be a sequence of semi-discrete solutions to (9). Then,
for any successive time steps, there exists a constant C > 0, such that

un+1 Hp ≤ C un Hp

A consequence of this regularity estimate is a bound on the error in the approx-

imation.
 
Lemma 2. Let f ∈ H p ∩ H01 and let the pair uh , V h ∈ V h solve
 ⎛ 2 ⎞2
     
uh  u τf 
= arg min Lτ − ⎝ 0 ⎠
Vh V  .
(u,V )∈V h  0 
8 J. H. Adler et al.

Let û be the exact solution of the corresponding PDE, i.e.,

2 2
−Δû + û = f in ∂Ω,
τ τ
û = 0 on ∂Ω.
Then,
 h  p
u − û 1 ≤ C(τ )h f ,
H H p−1
τ
where the constant C(τ ) may also depend on τ .
Proof. For a fixed τ , the PDE is a reaction-diffusion equation. Therefore, stan-
dard results from the FOSLS discretization of reaction-diffusion
  can be used
[8,9]. Note that for a standard FOSLS approach, C(τ ) = O τ12 , but a rescaling
of the equations may ameliorate this “worst-case scenario.”
Next, we make the following observation, which follows from the well-
posedness of the FOSLS formulation [8,9].
Lemma 3. Let (u1 , V1 ) ∈ V h and (u2 , V2 ) ∈ V h be two solutions to the following
FOSLS weak forms with different right-hand sides,
       
u1 u2
Lτ − F1 , Lτ φh = 0, Lτ − F2 , Lτ φh = 0 ∀φh ∈ V.
V1 V2
Then,
u1 − u2 H1 + V1 − V2 H1 ≤ C(τ ) F1 − F2 .
This, then, yields the following result.
Lemma 4. Given the solution to the semi-discrete Eq. (9), and the fully discrete
solution, we can bound the error in the L2 norm:
  C (τ )  
 h 
hp un H p + C2 (τ ) uhn − un  .
1
un+ 1 − un+ 12  ≤ (14)
2 τ

Proof. Let ũhn+ 1 be the scalar part of the FOSLS solution ũhn+ 1 , Ṽn+
h
1 of
2 2 2

2 2
− Δu + u = un , in Ω, u = 0 on ∂Ω, (15)
τ τ
where the exact semi-discrete solution un , at the previous time step, is used in
the right-hand side. By the triangle inequality,
     
 h     
un+ 1 − un+ 12  ≤ uhn+ 1 − ũhn+ 1  + ũhn+ 1 − un+ 12  . (16)
2 2 2 2

By Lemma 3, we have
     
 h     h h 
un+ 1 − ũhn+ 1  ≤ uhn+ 1 − ũhn+ 1  + Vn+ 1 − Ṽ
n+ 1
2 2 2 2 H1 2 2 H1 (17)
 
≤ C2 (τ ) uhn − un  .
 Energy Laws for FOSLS 9

The functions ũhn+ 1 and un+ 12 are, respectively, FOSLS and exact solutions
2
of the same boundary value problem (15). Hence, from Lemma 2, we have
  C(τ )
 h  C(τ ) p
ũn+ 1 − un+ 12  ≤ hp u n H p−1 ≤ h un Hp (18)
2 τ τ
Combining (16), (17) and (18), we obtain (14).

Finally, we have the following result on the approximation of the exact energy
law (13).
 h
un
Theorem 1. Let be the solution to the FOSLS system, (12), at time
Vnh
h h
step n (with un+ 1 and Vn+ 1 defined as before). There exists C(τ ) > 0 such that
2 2

⎛ ⎞ 
 h   uh 1 
 un+1 2 − uhn 2     n+ 2 
 h 2 2 h  ⎜ h ⎟ h 
 + Vn+ 1
 ≤ C(τ ) u n − u n min  V
⎝ n+ 1 ⎠ − Lτ φ .
2τ 2 τ φ h ∈V h  2 
 0 

Proof. To simplify the notation, define the energy law we wish to bound as

uhn+1 2
− uhn 2
Enh := h
+ Vn+ 1
2
.
2τ 2

Note that
 2  2  h   uh − uh 
1 uhn+1  − uhn  un+1 − uhn h n+ 12 n
h
= , un+ 1 = τ , un+ 1 ,
2 τ τ 2
2
2

and      
 h 2 h h h h h
Vn+ 1  = −∇ · Vn+ 1,u
n+ 1 + Vn+ 1 − ∇u
n+ 1,V
n+ 1 ,
2 2 2 2 2 2

where the latter equation is obtained by integration by parts, continuity of the

spaces, and appropriate boundary conditions. Thus,
 
uhn+ 1 − uhn  
h h h h h h
En = −∇ · Vn+ 1 + 2
τ , u n+ 1 + Vn+ 1 − ∇u n+ 1 , Vn+ 1
2 2 2 2 2
2
⎛ h ⎞
  h  ⎛ 2 h⎞ un+ 1 
un+ 1 τ u n
⎝ ⎠ ⎜ h 2⎟
= Lτ h
2
− 0 , ⎝Vn+ 1 ⎠ .
Vn+ 1 2
2 0 0

Using (12), for any φh ∈ V h ,

   ⎛2 ⎞ ⎛ h ⎞
h 
un+ 1
uhn+ 1 τ un ⎜ 2
⎟
Enh = Lτ h
2
− ⎝ 0 ⎠ , ⎝Vn+ 1 ⎠ − Lτ φ .
h h
Vn+ 1 2
2 0 0
10 J. H. Adler et al.

Next, consider adding and subtracting the solutions to the semi-discrete,

(11), and fully discrete, (12), FOSLS system from the previous time step,
  h  ⎛ 2 h⎞ ⎛ h ⎞ ⎛ h ⎞ 
u u − u n
un+ 1
un+ 1 τ n 2⎝ n ⎜ h 2⎟
h
En = Lτ h
2 − ⎝ 0 ⎠ + 0 ⎠ , ⎝Vn+ 1 ⎠ − Lτ φh
Vn+ 1 τ 2
2 0 0 0
⎛ h ⎞ ⎛ h ⎞ 
un − un un+ 1
2 ⎝ ⎜ 2
⎟
− 0 ⎠ , ⎝Vn+ h
1 ⎠ − Lτ φ
h
τ 2
0 0
  h  ⎛ ⎞ ⎛ h ⎞ 
u u 1
un+ 1 2 ⎝ n ⎠ ⎜ n+ h ⎟
2
h
= Lτ h
2 − 0 , ⎝Vn+ 1 ⎠ − Lτ φ
Vn+ τ 2
1
2 0 0
⎛ h ⎞ ⎛ h ⎞ 
un − un un+ 1
2 ⎝ ⎜ 2
⎟
− 0 ⎠ , ⎝Vn+ 1 ⎠ − Lτ φ
h h
τ 2
0 0
   ⎛ ⎞
 un 
 uhn+ 1 2  h 2 h h 

≤ Lτ 2 − ⎝ 0 ⎠ Mn + Mn un − un  ,
h
Vn+ 1 τ  τ
 2 0 
⎛ ⎞ 
 uh 1 
 n+ 2 
h ⎜ h ⎟ h
where we have defined Mn := min ⎝Vn+ 1 ⎠ − Lτ φ . Then, adding and
φ h ∈V h  2 
 0 
 
un+ 12
subtracting Lτ yields
Vn+ 12
   ⎛ ⎞ 0
 h   u :
n 
 u 1 − un+ 1 u n+ 2  ⎝
1 2   h 2 h h 
h 
En ≤ Lτ n+ 2 2
+ Lτ  − 0 ⎠ Mn + Mn un − un  .
h   
 Vn+ 1 − Vn+ 2  Vn+ 2 τ
0 
τ
1 1
2

Using the continuity of Lτ , followed by Lemma 3, gives

 
 uh 1 − u 1   
 n+  2
 + Mnh uhn − un 
n+
|Enh | ≤ C(τ )Mnh  h
2 2

 Vn+ 1 − Vn+ 1
2
 τ
2 H1
2   2  
≤ C(τ ) Mnh uhn − un  + Mnh uhn − un  .
τ τ
Combining the two terms completes the proof.
To provide a better bound for the FOSLS energy law (13), we introduce a mea-
sure for the truncation error defined as
⎛ ⎞ 
 uh 1 (v) 
 n+ 
1 ⎜ h 2 ⎟ h
δn = max min  V
⎝ n+ 1 ⎠(v) − Lφ , (19)
v∈H p+1 (Ω) v H p+1 φ h ∈V h  2 
 0 
 Energy Laws for FOSLS 11

where uhn+ 1 (v) and Vn+

h
1 (v) are the corresponding solutions to the fully discrete
2 2
problem with u0 = v as the initial condition.
Corollary 1. Using the same assumptions as Theorem 1 and Assumption 1,
 h 
 un+1 2 − uhn 2  C(τ )δ p
 V h 2
 ≤ τ h u0 H p+1 , δ = max
2
 + n+ 1 δn . (20)
2τ 2 n

Proof. Using the definitions of uhn+ 1 and un+ 12 , the triangle inequality, and
2
Lemma 4,

uhn+1 − un+1 ≤ 2 uhn+ 1 − un+ 12 + uhn − un

2

2
≤ C1 (τ )hp un Hp + (C2 (τ ) + 1) uhn − un .
τ
An induction argument then gives
n
2 j−1 n
uhn − un ≤ C1 (τ )h p
(C2 (τ ) + 1) un−j Hp + (C2 (τ ) + 1) uh0 − u0 .
τ j=1

With Assumption 1,
n
2 j−1 n
uhn − un ≤ C1 (τ )hp (C2 (τ ) + 1) un−j Hp + (C2 (τ ) + 1) u0 H p+1 .
τ j=1

Using some regularity arguments for each ui , we get,

uhn − un ≤ C(n)hp u0 H p+1 .

Then, with the definition of δ and the result from Theorem 1, the proof is com-
plete.
We note that the bound in Corollary 1 is a rather pessimistic one. At a
fixed time, t, we expect the quality of both the fully discrete and semi-discrete
approximations to the true solution to improve as τ → 0 and more time-steps
are used to reach time t; thus, uhn − un should decrease as τ → 0 for n = t/τ .
Furthermore, for the unforced heat equation, we expect both uhn and un to
decrease in magnitude with n, but this is not accounted for in the bound in
Corollary 1. The bound above worsens with smaller τ and bigger n, showing
the limitations of bounding uhn − un by terms depending only on u0 and the
finite-element space.
Remark 2. As shown in the numerical experiments, Sect. 5, the constant δ
defined in (19) is of order hp for a smooth
  solution. This indicates that the
energy law (13) holds with order O h2p . While the theoretical justification
of such statement may be plausible, it is nontrivial as the discrete quantities
involved in the definition of δ do not possess enough regularity (they are just
finite-element functions, only in H 1 ).
12 J. H. Adler et al.

4 Stokes’ Equations
Next, we return to the time-dependent Stokes’ equations, (3)–(4). For simplicity,
we again assume ν = 1, and rewrite the equations using Dirichlet boundary
conditions for the normal components of the velocity field, and zero-mean average
for the pressure field,
∂u(x, t)
− Δu(x, t) + ∇p(x, t) = 0 ∀x ∈ Ω, ∀t > 0 (21)
∂t
∇ · u(x, t) = 0 ∀x ∈ Ω, ∀t > 0 (22)
n · u(x, t) = 0 ∀x ∈ ∂Ω, ∀t ≥ 0 (23)
u(x, 0) = g(x) ∀x ∈ Ω̄, (24)

p(x, t) dV = 0 ∀t ≥ 0. (25)
Ω

Using a similar semi-discretization in time with Crank-Nicolson that was done

in (9) yields
un+ 12 − un
τ  − Δun+ 12 + ∇pn+ 12 = 0,
2
∇ · un+ 12 = 0,
n · un+ 12 (x) = 0 ∀x ∈ ∂Ω,
 (26)
pn+ 12 dV = 0 ∀n ≥ 0,
Ω
un+1 = 2un+ 12 − un ,
pn+ 12 = 2pn+ 12 − pn .

To use the FOSLS method, we put the operator into a first-order system in a
similar fashion to the heat equation. Least-squares formulations are well-studied
for Stokes’ system and we consider a simple, velocity-gradient-pressure formu-
lation, where a new gradient tensor, V = ∇u, is used to obtain an H 1 -elliptic
equivalent system [4,6,14]:
⎛ ⎞ ⎛ ⎞
−∇ · Vn+ 12 + ∇pn+ 12 + τ2 un+ 12 2
un
⎛ ⎞ ⎜ ⎟ ⎜ τ
un+ 12 ⎜ ∇ · un+ 12 ⎟ ⎜ 0 ⎟ ⎟
⎜ ⎟ ⎜
Lτ ⎝Vn+ 12 ⎠ = ⎜ Vn+ 12 − ∇un+ 12 ⎟=⎜ 0 ⎟ ⎟. (27)
⎜ ⎟ ⎝
pn+ 12 ⎝ ∇ × Vn+ 12 ⎠ 0 ⎠
∇trVn+ 1 0
2

Appropriate boundary conditions on the continuous solution, such as n · u = 0,

gives rise to tangential boundary conditions on V, V × n = 0, where n is
the normal vector to the boundary. Ultimately, the corresponding semi-discrete
energy law is
||un+1 ||2 − ||un ||2
= −||Vn+ 12 ||2 . (28)
2τ
 Energy Laws for FOSLS 13

Finally, we minimize the residual of (27) over a finite-dimensional subspace of

the product H 1 Sobolev space in the L2 norm obtaining the weak equations,
⎛ 2 h⎞
⎛ h ⎞ τ un
 un+ 1 ⎜ 0 ⎟ 
⎜V h ⎟ ⎜
2 ⎜ ⎟
Lτ ⎝ n+ 12 ⎠ − ⎜ 0 ⎟ ⎟ , Lτ φ
h
= 0 ∀φh ∈ V h . (29)
h
pn+ 1 ⎝ 0 ⎠
2
0

Note that the weak system is similar to (12) and the energy law is identical
to (13) in vector form. Thus, all the above theory still holds subject to enough
regularity of the solution to the time-dependent Stokes’ equations [21,22] and a
suitable generalization of the definition of δ.

5 Numerical Experiments
For the numerical results presented here, we use a C++ implementation of the
FOSLS algorithm, using the modular finite-element library MFEM [20] for man-
aging the discretization, mesh, and timestepping. The linear systems are solved
by direct method using the UMFPACK package [10].

5.1 Heat Equation

First, we consider the heat Eq. (6), and its discrete FOSLS formulation, (12),
on a triangulation of Ω = (0, 1) × (0, 1). The data is chosen so that the true
2
solution is u(x, y, t) = sin(πx) sin(πy)e−2π t . Note that this solution satisfies the
boundary conditions and other assumptions discussed above.

103
p=1
p=2
100 p=3
h2 convergence
Energy Law Error

h4 convergence
10−3
h6 convergence

10−6

10−9

10−12
0 2 4 6
# Refinement Levels,

Fig. 1. Energy law error, (13), vs. number of mesh refinements,  (h = 21 ), for the
FOSLS discretization of the heat equation, (12), using various orders of the finite-
element space (p = 1 - linear; p = 2 - quadratic; and p = 3 - cubic). One time step is
performed with τ = 0.005.
14 J. H. Adler et al.

Figure 1 displays the convergence of the energy law to zero as the mesh is
refined for a fixed time step. The convergence is O h2p , where p is the order of
the finite-element space being considered, confirming Theorem 1. It also suggests
that the constant δ is O (hp ), as is remarked above.

100
10−1 p=1
p=2
p=3
10−4
10−4
Energy Law Error

Energy Law Error

10−8 10−7

10−12 10−10

10−13
10−16
0 20 40 60 0 0.2 0.4 0.6 0.8 1
# Time Steps, n Time Step Size, τ

(a) (b)

Fig. 2. Energy law error, (13), vs. (a) number of time steps, n (with fixed τ = 0.005),
and (b) time step size, τ , for the FOSLS discretization of the heat equation, (12), using
various orders of the finite-element space (p = 1 - linear; p = 2 - quadratic; and p =
1
3 - cubic). Mesh spacing is h = 32 .

Figure 2 indicates how the timestepping affects the convergence of the energy
law. As discussed above, taking more time steps decreases the error in the energy
law, showing that we can improve the results on the bound, un − uhn . On the
other hand, if only one time step is taken, the convergence slightly worsens
for small τ , which is consistent with the constants found in Theorem 1 and
Corollary 1.

5.2 Stokes’ Equations

Next, we consider Stokes’ Equations, (21), and the FOSLS discretization

described above, (29). The same domain, Ω = (0, 1) × (0, 1), is used, and we
assume data that yields the exact solution,
 
sin(πx) cos(πy) 2
u(x, t) = e−2π t ,
− cos(πx) sin(πy)
p(x, t) = 0.

This produces a C ∞ solution that satisfies the appropriate boundary conditions

and regularity arguments needed for the bounds on the energy law described
above.
Another random document with
no related content on Scribd:
 "We are a people of peaceful traders—shopkeepers, our rivals
of the Continent affirm—and are consequently at war on only
eight points of the globe, with forces which in the aggregate
only just exceed sixty thousand men. There are thirty-five
thousand on the Indian Frontier fighting the clansmen of the
Northern Himalayas, who, according to the Afridi sub-officers
interrogated by Sir Henry Havelock-Allan, are all eager to
enter our service; twenty-five thousand about to defeat the
Khalifa at Omdurman; a thousand doing sentry duty in Crete;
four hundred putting down an outbreak in Mekran; three hundred
crushing a mutiny in Uganda; and some hundreds more restoring
order in Lagos, Borneo, and Basutoland. All these troops,
though of different nationalities—Englishmen, Sikhs, Ghoorkas,
Rajpoots, Malays, Egyptians, Soudanese, Haussas, and Wagandas—
are under British officers, are paid from funds under British
control, and are engaged in the self-same work, that of
solidifying the 'Pax Britannica,' so that a commercial
civilisation may have a fair chance to grow."

The Spectator (London), February 5, 1898.

ENGLAND: A. D. 1898 (February).

Resentment shown to China for rejection of a loan,
through Russian influence.
Chinese agreement not to alienate the Yang-tsze region
and to open internal waters to steam navigation.

See (in this volume)

CHINA: A. D. 1898 (FEBRUARY).

ENGLAND: A. D. 1898 (February-May).

Native revolt in the Sierra Leone Protectorate.

See (in this volume)

SIERRA LEONE PROTECTORATE.
ENGLAND: A. D. 1898 (March-April).
Unsuccessful opposition to Russian lease of Port Arthur
and Talienwan from China.
Compensatory British lease of Wei-hai Wei.

See (in this volume)

CHINA: A. D. 1898 (MARCH-JULY).

ENGLAND: A. D. 1898 (April-August).

Further exactions from China.
Lease of territory opposite Hong Kong, etc.

See (in this volume)

CHINA: A. D. 1898 (APRIL-AUGUST).

ENGLAND: A. D. 1898 (May).

Death of Mr. Gladstone.

After a long and painful illness, the great statesman and

leader of the Liberal party in England, William Ewart
Gladstone, died on the 19th of May. His death drew tributes in
Parliament from his political opponents which exalted him quite
to the height of great distinction that those who followed him
would claim. It was said by Lord Salisbury that "the most
distinguished political name of the century had been withdrawn
from the roll of Englishmen." Mr. Balfour described him as
"the greatest member of the greatest deliberative assembly
that the world had yet seen": and expressed the belief that
"they would never again have in that assembly any man who
could reproduce what Mr. Gladstone was to his contemporaries."

Lord Rosebery paid an eloquent tribute to the dead statesman.

"This country." he said, "this nation, loves brave men. Mr.
Gladstone was the bravest of the brave. There was no cause so
hopeless that he was afraid to undertake it; there was no
amount of opposition that would cowe him when once he had
undertaken it. My lords, Mr. Gladstone always expressed a hope
 that there might be an interval left to him between the end of
his political and of his natural life. That period was given
to him, for it is more than four years since he quitted the
sphere of politics. Those four years have been with him a
special preparation for his death, but have they not also been
a preparation for his death with the nation at large?
{210}
Had he died in the plenitude of his power as Prime Minister,
would it have been possible for a vigorous and convinced
Opposition to allow to pass to him, without a word of dissent,
the honours which are now universally conceded? Hushed for the
moment are the voices of criticism, hushed are the controversies
in which he took part; hushed for the moment is the very sound
of party faction. I venture to think that this is a notable
fact in our history. It was not so with the elder Pitt. It was
not so with the younger Pitt. It was not so with the elder
Pitt, in spite of his tragic end, of his unrivalled services,
and of his enfeebled old age. It was not so with the younger
Pitt, in spite of his long control of the country and his
absolute and absorbed devotion to the State. I think that we
should remember this as creditable not merely to the man, but
to the nation." With the consent of Mrs. Gladstone and family,
a public funeral was voted by Parliament, and the remains of the
great leader were laid, with simple but impressive ceremonies,
in Westminster Abbey, on the 28th of May.

ENGLAND: A. D. 1898 (June).

The Sugar Conference at Brussels.

See (in this volume)

SUGAR BOUNTIES.

ENGLAND: A. D. 1898 (July).

The Local Government Act for Ireland.

See (in this volume)

IRELAND: A. D. 1898 (JULY).
ENGLAND: A. D. 1898 (July-December).
In the Chinese "Battle of Concessions."

See (in this volume)

CHINA: A. D. 1898 (FEBRUARY-DECEMBER).

ENGLAND: A. D. 1898 (September-November).-

The Nile question with France.
Marchand's expedition at Fashoda.

See (in this volume)

EGYPT: A. D. 1898 (SEPTEMBER-NOVEMBER).

ENGLAND: A. D. 1898 (December).

Imperial Penny Postage.

On Christmas Day, 1898, the Imperial penny postage came into

operation,—i. e., it became possible to send for a penny a
letter not above half an ounce in weight to all places in the
British Empire, except the Australasian Colonies and the Cape.
"Thousands of small orders and business transactions and
millions of questions and answers will fly round the world at
a penny which were too heavily weighted at two-pence
halfpenny. The political effect of the fact that it will not
now be necessary to think whether an address is outside the
United Kingdom, but only whether it is inside the British
Empire, will be by no means insignificant. If people will only
let the Empire alone we shall ultimately weave out of many
varied strands—some thick, some thin—a rope to join the
Motherland and the Daughter States which none will be able to
break. Not an unimportant thread in the hawser will
be,—letters for a penny wherever the Union Jack is flown."

The Spectator (London),

December 31, 1898.
ENGLAND: A. D. 1898-1899.
Joint High Commission for settlement of pending questions
between the United States and Canada.

See (in this volume)

CANADA: A. D. 1898-1899.

ENGLAND: A. D. 1898-1899 (June-June).

Convention with France defining West African and
Sudan possessions.

See (in this volume)

NIGERIA: A. D. 1882-1899.

ENGLAND: A. D. 1899.
Dealings with anti-missionary demonstrations in China.

See (in this volume)

CHINA: A. D. 1899.

ENGLAND: A. D. 1899 (January).

Agreement with Egypt, establishing the Anglo-Egyptian
Condominium in the Sudan.

See (in this volume)

EGYPT: A. D. 1899 (JANUARY).

ENGLAND: A. D. 1899 (March-April).

Agreement with Russia concerning railway interests in China.

See (in this volume)

CHINA: A. D. 1899 (MARCH-APRIL).

ENGLAND: A. D. 1899 (May-June).

The Bloemfontein Conference with President Kruger.

See (in this volume)

 SOUTH AFRICA (THE TRANSVAAL): A. D. 1899 (MAY-JUNE).

ENGLAND: A. D. 1899 (May-July).

Representation in the Peace Conference at The Hague.

See (in this volume)

PEACE CONFERENCE.

ENGLAND: A. D. 1899 (June-October).

Arbitration and settlement of the Venezuela boundary question.

See (in this volume)

VENEZUELA: A. D. 1896-1899.

ENGLAND: A. D. 1899 (July).

Passage of the London Government Act.

See (in this volume)

LONDON: A. D. 1899.

ENGLAND: A. D. 1899 (July-September).

Discussion of proposed amendments to the Franchise Law
of the South African Republic.

See (in this volume)

SOUTH AFRICA (THE TRANSVAAL):
A. D. 1899 (JULY-SEPTEMBER).

ENGLAND: A. D. 1899 (August).

The Board of Education Act.

An Act of Parliament which became law on the 9th of August,

1899, and operative on the 1st of April, 1900, created a
national Board of Education, "charged with the superintendence
of matters relating to education in England and Wales," and
taking the place of the Committee of the Privy Council on
Education, by which that function had previously been
 performed. The Act provided that the Board "shall consist of a
President, and of the Lord President of the Council (unless he
is appointed President of the Board), Her Majesty's Principal
Secretaries of State, the First Commissioner of Her Majesty's
Treasury, and the Chancellor of Her Majesty's Exchequer. … The
President of the Board shall be appointed by Her Majesty, and
shall hold office during Her Majesty's pleasure." The Act
provided further for the creation by Her Majesty in Council of
"a Consultative Committee consisting, as to not less than
two-thirds, of persons qualified to represent the views of
Universities and other bodies interested in education, for the
purpose of—(a) framing, with the approval of the Board of
Education, regulations for a register of teachers, … with an
entry in respect to each teacher showing the date of his
registration, and giving a brief record of his qualifications
and experience; and (b) advising the Board of Education on any
matter referred to the committee by the Board."

62 & 63 Victoria, chapter 33.

ENGLAND: A. D. 1899 (August).

Instructions to the Governor of Jamaica.

See (in this volume) JAMAICA: A. D. 1899.

ENGLAND: A. D. 1899 (September-October).

Preparations for war in South Africa.
The Boer Ultimatum.

See (in this volume)

SOUTH AFRICA (THE TRANSVAAL AND ORANGE FREE STATE):
A. D. 1899 (SEPTEMBER-OCTOBER).

ENGLAND: A. D. 1899 (October-November).

Opening circumstances of the war in South Africa.
Want of preparation.
 See (in this volume)
SOUTH AFRICA (THE FIELD OF WAR):
A. D. 1899 (OCTOBER-NOVEMBER).

ENGLAND: A. D. 1899 (November).

Adhesion to the arrangement of an "open door" commercial
policy in China.

See (in this volume)

CHINA: A. D. 1899-1900 (SEPTEMBER-FEBRUARY).

{211}

ENGLAND: A. D. 1899 (November).

Withdrawal from the Samoan Islands, with compensations in the
Tonga and Solomon Islands and in Africa.

See (in this volume)

SAMOAN ISLANDS.

ENGLAND: A. D. 1899-1900.
Renewed investigation of the Old-Age Pension question.

On the initiative of the government, a fresh investigation of

the question of old-age pensions was opened in 1899 by a
select committee of the House of Commons, under the
chairmanship of Mr. Chaplin. The report of the Committee, made
in the following year, suggested the following plan: Any
person, aged 65, whether man or woman, who satisfied the
pension authority that he or she"

(1) Is a British subject;

(2) Is 65 years of age;

(3) Has not within the last 20 years been convicted of an

offence and sentenced to penal servitude or imprisonment
without the option of a fine;

(4) Has not received poor relief, other than medical relief,
unless under circumstances of a wholly exceptional character,
during twenty years prior to the application for a pension;

(5) Is resident within the district of the pension authority;

(6) Has not an income from any source of more than 10s. a
week; and

(7) Has endeavoured to the best of his ability, by his

industry or by the exercise of reasonable providence, to make
provision for himself and those immediately dependent on
him—"should receive a certificate to that effect and be
entitled to a pension. The amount of pension to be from 5s. to
7s. a week.

As a means of ascertaining approximately the number of persons

in the United Kingdom who would be pensionable under this
scheme, a test census was taken in certain districts made as
representative as possible by the inclusion of various kinds
of population. In each of the selected areas in Great Britain
a house-to-house visitation was made with a view of
ascertaining how many of the aged would satisfy the conditions
of the scheme. In Ireland a similar census had to be abandoned
as impracticable because "the officials, although they
proceeded courteously, were received with abuse"; but the Poor
Law inspectors framed some rough estimates after consultation
with local authorities. Altogether the inquiry in Great
Britain extended to a population of rather over half a million
persons. From facts thus obtained the following estimate of
the cost of the proposed pensioning project was deduced:

Estimated number of persons

over 65 years of age in 1901
2,016,000
Deduct:
1. For those whose incomes exceed 10s. a week
741,000
2. For paupers
515,000
3. For aliens, criminals, and lunatics
32,000
4. For inability to comply with thrift test
72,700

Total deductions
1,360,700

Estimated number of pensionable persons

655,000

Estimated cost (the average pension being

taken at 6s. a week)
£9,976,000
Add administrative expenses (3 per cent.)
£299,000

Total estimated cost.

£10,275,000

In round figures.
£10,300,000

The Committee estimated, still further, that the cost would

rise to £15,650,000 by 1921. No legislative action was taken
on the report.
ENGLAND: A. D. 1899-1900 (October-January).
Troops from Canada for the South African War.

See (in this volume)

CANADA: A. D. 1899-1900.

ENGLAND: A. D. 1899-1901.
The Newfoundland French Shore question.

See (in this volume)

NEWFOUNDLAND: A. D. 1899-1901.

ENGLAND: A. D. 1900.
Industrial combinations.

See (in this volume)

TRUSTS: IN ENGLAND.

ENGLAND: A. D. 1900.
Naval strength.

See (in this volume)

NAVIES OF THE SEA POWERS.

ENGLAND: A. D. 1900 (January-March).

The outbreak of the "Boxers" in northern China.

See (in this volume)

CHINA: A. D. 1900 (JANUARY-MARCH).

ENGLAND: A. D. 1900 (February).

Compulsory education.

A bill introduced in Parliament by a private member,

unsupported by the government, providing that the earliest
date at which a child should be permitted to leave school
 should be raised from 11 to 12 years, was passed, only one
member of the Cabinet voting for it.

ENGLAND: A. D. 1900 (February).

Negotiation of a convention with the United States relative
to the projected Interoceanic Canal.

See (in this volume)

CANAL, INTEROCEANIC: A. D. 1900 (DECEMBER).

ENGLAND: A. D. 1900 (March).

Overtures of peace from the Boer Presidents.
Reply of Lord Salisbury.

See (in this volume)

SOUTH AFRICA (THE FIELD OF WAR):
A. D. 1900 (MARCH).

ENGLAND: A. D. 1900 (May).

Annexation of Orange Free State by right of conquest.

See (in this volume)

SOUTH AFRICA (ORANGE FREE STATE): A. D. 1900 (MAY).

ENGLAND: A. D. 1900 (June-December).

Co-operation with the Powers in China.

See (in this volume)

CHINA.

ENGLAND: A. D. 1900 (July).

Passage of the "Commonwealth of Australia Constitution Act,"
federating the Australian Colonies.

See (in this volume)

AUSTRALIA: A. D. 1900;
and CONSTITUTION OF AUSTRALIA.
ENGLAND: A. D. 1900 (September).
Proclamation of the Commonwealth of Australia.

See (in this volume)

AUSTRALIA: A. D. 1900 (SEPTEMBER-DECEMBER).

ENGLAND: A. D. 1900 (September-October).

Dissolution of Parliament.
Election of a new Parliament.
Victory for the Conservatives and Liberal Unionists.

By royal proclamation, September 17, the existing Parliament

was dissolved and order given for the issue of writs calling a
new Parliament, the elections for which were held in October,
concluding on the 24th of that month. The state of parties in
the House of Commons resulting from the election was as
follows: Conservatives, 334, Liberal Unionists, 68; total
supporters of the Unionist Ministry, 402. Liberals and Labor
members, 186, Nationalists (Irish), 82; total opposition, 268.
Unionist majority, 134, against 128 in the preceding
Parliament. The issues in the election were those growing out
of the South African War. Although most of the Liberals upheld
the war, and the annexation of the South African republics,
they sharply criticised the prior dealings of the Colonial
Secretary, Mr. Chamberlain, with the Transvaal Boers, and the
general conduct of the war. A number of the leading Liberals
were uncompromising in condemnation of the war, of the policy
which caused it, and of the proposed extinction of Boer
independence. The sentiment of the country was shown by the
election to be strongly against all questioning of the
righteousness of the war or of the use to be made of victory
in it.

{212}

ENGLAND: A. D. 1900 (October).

 Anglo-German agreement concerning policy in China.

See (in this volume)

CHINA: A. D. 1900 (AUGUST-DECEMBER).

ENGLAND: A. D. 1900 (October).

Annexation of the Transvaal.

See (in this volume)

SOUTH AFRICA (THE TRANSVAAL):
A. D. 1900 (OCTOBER).

ENGLAND: A. D. 1900 (November-December).

The Fourth Ministry of Lord Salisbury.
Brief session of Parliament.

For the fourth time, Lord Salisbury was called to the lead in
government, and formed his Ministry anew, making considerable
changes. He relieved himself of the conduct of Foreign Affairs
(which was transferred to the Marquis of Lansdowne), and took,
with the office of Prime Minister, that of Lord Privy Seal. Mr.
Brodrick, who had been an Under Secretary, succeeded Lord
Lansdowne as Secretary of State for War. Mr. Balfour continued
to be First Lord of the Treasury, and Leader of the House; Mr.
Chamberlain remained in the Colonial Office. Mr. Goschen
retired.

Parliament met on the 6th of December, for the purpose set

forth in a remarkably brief "Queen's Speech," as follows: "My
Lords, and Gentlemen, It has become necessary to make further
provision for the expenses incurred by the operations of my
armies in South Africa and China. I have summoned you to hold
a Special Session in order that you may give your sanction to
the enactments required for this purpose. I will not enter
upon other public matters requiring your attention until the
ordinary meeting of Parliament in the spring." The estimates
of the War Office called for £16,000,000, and it was voted
 after a few days of debate, in which the causes and conduct of
the war were criticised and defended by the two parties, and,
on the 15th, Parliament was prorogued to the 14th of February,
1901, by the Queen's command.

ENGLAND: A. D. 1900 (December).

Fall of stones at Stonehenge.

See (in this volume)

STONEHENGE.

ENGLAND: A. D. 1900 (December).

Parliamentary statements of the number of men employed in the
South African War, and the number dead and disabled.

In the House of Commons, December 11, Mr. Brodrick, Secretary

of State for War, moved a vote of £16,000,000, required for
the current year, to meet additional expenditure in South
Africa and China. In the course of his remarks, explanatory of
the need for this supplementary supply, he made the following
statement: "When the war broke out we had in South Africa in
round figures 10,000 men, all Regular troops. We have in the
14 months' which have since elapsed sent from this country and
landed in South Africa 175,000 Regular soldiers, a number which
exceeds by far any number which any Minister from this bench
or any gentleman sitting behind these benches or in front of
them ever suggested that this country ought to be in a
position to ship to any part of the world, and a number far in
excess of that which during any period that I have sat in the
house any member of the House, except an official, would have
been willing to believe that the War Office could find to
dispose of. But they are not the only troops. We have called
on them, I will not say to the extreme limit of our power,
but, at all events, with an unsparing hand. But you have in
addition, as this return will show, some 40,000 Volunteers of
various descriptions from the United Kingdom—40,000 including
the Imperial Yeomanry, whose service is spoken of by every
officer under whom they have served with such satisfaction; 30
Militia regiments, who are also Volunteers, since their term
of service was only for the United Kingdom and who have gone
abroad at great personal sacrifice to themselves; and the
volunteer companies who have joined the Regular battalions.
You have also got 40,000 colonial troops, to a large extent,
no doubt, men raised in the colonies affected, and as
everybody knows to a still larger extent consisting of men who
have gone for a year from Australia, Canada, and other
places."

Sir William Harcourt replied to Mr. Brodrick, not in

opposition to the motion, but in criticism of the conduct of
the war. Referring to a return submitted by the War Office, he
analyzed its showing of facts, thus: "Now just let us look at
this table. By some accident it only gives the rank and file
and non-commissioned officers. It is a very terrible return,
and I think it is worthy of the attention of the men who
delight in war, of whom, I am afraid, there are unhappily not
a few. I have made a short analysis of the paper. It shows
that the garrison at the Cape before the war was 9,600.
Reinforcements of 6,300 men were sent out in October last year
and from India 5,600, which with the former garrison made up
21,000 in all when the war broke out. Up to August, that is,
after the last estimate for 1900, according to this table
267,000 men had been in arms in South Africa—that is without
the officers. Therefore I will call it 270,000 men in round
numbers. I think the right honourable gentleman made a mistake
when he said that the colonial troops were more numerous from
beyond the seas than they were in the Cape. This return shows
that the men raised in South Africa were 30,000, and, apart
from them, the colonials from beyond the seas were 11,000.
According to the last return there were 210,000 men in South
Africa. You will observe there is a balance of some 60,000 or
70,000 men. What has become of those men? You would find from
this return, one would suppose, that a good many of these have
returned safe and sound to England. No, Sir; the men who have
 returned to England according to this paper, not invalids, are
7,500 and to the colonies 3,000 more. That makes 10,000 men,
or with the officers about 11,000 men. But since July you have
sent out 13,000 men to South Africa, more, in fact, than you
have been bringing home, and yet you have only 210,000 men
there. Now, Sir, how is this accounted for? First of all you
have the heading, 'killed or died of wounds,' 11,000 men. You
have 'wounded,' 13,000, you have 'in hospital in South
Africa,' 12,000, and you have 'returned to England, sick,
wounded, or died on passage,' 36,000 men. That is the balance.
Seventy thousand men have been killed, wounded, or disabled,
or have died in this war. And now what is the prospect that is
held before us with this force, once 270,000 men, and now
210,000, in South Africa? Lord Roberts has declared that the
war is over, yet you hold out to us no prospect of diminishing
the force you have in South Africa of 210,000 men."

{213}

ENGLAND: A. D. 1901 (January).

Death of Queen Victoria.

The following notice, which appeared in the "Court Circular,"

on the 18th of January, dated from the winter residence of the
Queen at Osborne House, in the Isle of Wight, seems to have
been the first intimation to the country of its sovereign's
failing health: "The Queen has not lately been in her usual
health and is unable for the present to take her customary
drives. The Queen during the past year has had a great strain
upon her powers, which has rather told upon her Majesty's
nervous system. It has, therefore, been thought advisable by
her Majesty's physicians that the Queen should be kept
perfectly quiet in the house and should abstain for the
present from transacting business." It was subsequently found,
as stated in an "authoritative account" by the "British
Medical Journal," and the "Lancet," that "the Queen's health
for the past 12 months had been failing, with symptoms mainly
of a dyspeptic kind, accompanied by impaired general
nutrition, periods of insomnia, and later by occasional slight
and transitory attacks of aphasia, the latter suggesting that
the cerebral vessels had become damaged, although her
Majesty's general arterial system showed remarkably few signs
of age. … The dyspepsia which tended to lower her Majesty's
original robust constitution was especially marked during her
last visit to Balmoral. It was there that the Queen first
manifested distinct symptoms of brain fatigue and lost notably
in weight. These symptoms continued at Windsor, where in November
and December, 1900, slight aphasic symptoms were first
observed, always of an ephemeral kind, and unattended by any
motor paralysis. … A few days before the final illness
transient but recurring symptoms of apathy and somnolence,
with aphasic indications and increasing feebleness, gave great
uneasiness to her physician." Before the publication of the
cautious announcement quoted above, the symptoms had become
too grave to leave any doubt as to the near approach of death.
It came on Tuesday, the 22d of January, at half past six
o'clock in the evening, the dying Queen being then surrounded
by a large number of her many children, grandchildren and
great grandchildren, whom she recognized, it is said, within a
few moments of the end. The eldest of the Queen's children,
the Empress Frederick, was kept from her mother's side a this
last hour by serious illness of her own; but the Emperor
William, of Germany (son of the Empress Frederick and eldest
grandson of Queen Victoria) had hastened to the scene and
showed a filial affection which touched English hearts.

On Friday, the first day of February, the remains of the Queen

were borne from the island where she died to Portsmouth, between
long lines of battle-ships and cruisers—British, German,
French, Italian, Japanese, Belgian and Portuguese. The scene
of the funeral voyage was impressively described by a
correspondent of the New York "Sun," as follows: "Nature was
never kindlier. The smiling waters of the Solent were as calm
as on a summer's morning. It was 'Queen's weather' to the very
last. The cavalcade which wended slowly through the narrow
lane, green even in midwinter, down through the streets of the
little town of Cowes to the Trinity pier was a funeral
procession such as the world had never seen before. Kings and
princes, a Queen and princesses, walked humbly between black
lines of mourning islanders, escorting the coffin of the dead
sovereign. Then followed a sight far more notable and more
impressive, indeed, than the great tribute the great capital
of the empire will pay to-morrow. It was the transit of the
funeral yacht across the waters between lines of steel which
are England's bulwarks against the world. Battleship after
battleship thundered its grief, band after band wailed its
dirge and crew after crew bowed low their heads as the pigmy
yacht swept past, bearing no passengers save an admiral on the
bridge and four red-coated guards at the corners of the
simple, glowing white bier resting amidships. It was a picture
neither a painter's brush nor an orator's eloquence could
reproduce. … The boat slowly glided on in the mellow light of
the afternoon sun, herself almost golden in hue, sharply
contrasting with the black warships. The ears also were
assailed in strange contrast, the sad strains of Beethoven's
funeral march floating over the water being punctuated by the
roar of minute guns from each ship. Somehow it was not
incongruous and one felt that it was all a great and majestic
tribute to a reign which was an era and to a sovereign to whom
the world pays its highest honors."

On the following day the remains were conveyed by railway from

Portsmouth to London, carried in solemn procession through the
streets of the capital, and thence by railway to Windsor,
where the last rites were performed on Monday, the 4th. The
Queen was then laid to rest, by the side of her husband, in
the mausoleum which she had built at Frogmore.

Of the sincerity with which Queen Victoria had been loved by

her own people and respected and admired by the world at
large, and of the genuineness of sorrow that was manifested
 everywhere at her death, there can be no doubt. To the
impressiveness of the ending of an unexampled period of
history there was added a true sense of loss, from the
disappearance of a greatly important personage, whose high
example had been pure and whose large influence had been good.

Among all the tributes to the Queen that were called out by
her death none seem so significant and so fully drawn from
knowledge of what she was in her regal character, as the words
that were spoken by Lord Salisbury in the House of Lords, at
the meeting of Parliament on the Friday following her death.
"My lords." he said, "the late Queen had so many titles to our
admiration that it would occupy an enormous time to glance at
them even perfunctorily; but that on which I think your
lordships should most reflect, and which will chiefly attach
to her character in history, is that, being a constitutional
monarch with restricted powers, she reigned by sheer force of
character, by the lovableness of her disposition, over the
hearts of her subjects, and exercised an influence in moulding
their character and destiny which she could not have done more
if she had bad the most despotic power. She has been a great
instance of government by example, by esteem, by love; and it
will never be forgotten how much she has done for the
elevation of her people, not by the exercise of any
prerogative, not by the giving of any commands, but by the
simple recognition and contemplation of the brilliant
qualities which she has exhibited in her exalted position. My
lords, it may be, perhaps, proper that those who, like noble
lords opposite and myself, have had the opportunity of seeing
the close workings of her character in the discharge of her
duties as Sovereign, should take this opportunity of
testifying to the great admiration she inspired and the great
force which her distinguishing characteristics exercised over
all who came near her.
{214}
The position of a Constitutional Sovereign is not an easy one.
Duties have to be reconciled which sometimes seem far apart.