Modeling, Simulation and Optimization of Complex Processes HPSC 2018: Proceedings of the 7th International Conference on High Performance Scientific Computing, Hanoi, Vietnam, March 19-23, 2018

402 Pages·2021·12.289 MB·English

by  Hans Georg Bock, Willi Jäger, Ekaterina Kostina, Hoang Xuan Phu

Hans Georg Bock · Willi Jäger ·
Ekaterina Kostina · Hoang Xuan Phu Editors
Modeling, Simulation
and Optimization
of Complex Processes
HPSC 2018
Modeling, Simulation and Optimization of Complex
Processes HPSC 2018
ä
Hans Georg Bock Willi J ger
(cid:129) (cid:129)
Ekaterina Kostina Hoang Xuan Phu
(cid:129)
Editors
Modeling, Simulation
and Optimization of Complex
Processes HPSC 2018
Proceedings of the 7th International
fi
Conference on High Performance Scienti c
–
Computing, Hanoi, Vietnam, March 19 23,
2018
123
Editors
Hans GeorgBock Willi Jäger
Interdisciplinary Center Interdisciplinary Center
for ScientificComputing (IWR) for ScientificComputing (IWR)
Heidelberg University Heidelberg University
Heidelberg, Germany Heidelberg, Germany
Ekaterina Kostina Hoang XuanPhu
Interdisciplinary Center Institute of Mathematics
for ScientificComputing (IWR) Vietnam Academy of Science
Heidelberg University andTechnology
Heidelberg, Germany Hanoi, Vietnam
ISBN978-3-030-55239-8 ISBN978-3-030-55240-4 (eBook)
https://doi.org/10.1007/978-3-030-55240-4
Mathematics Subject Classification: 34B15, 35Q35, 35Q92, 49K15, 49J15, 49M30, 65K05, 65L05,
70E60,93B30,93B40
©SpringerNatureSwitzerlandAG2021
Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpart
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
orinformationstorageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilar
methodologynowknownorhereafterdeveloped.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publicationdoesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfrom
therelevantprotectivelawsandregulationsandthereforefreeforgeneraluse.
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, expressed or implied, with respect to the material contained
hereinorforanyerrorsoromissionsthatmayhavebeenmade.Thepublisherremainsneutralwithregard
tojurisdictionalclaimsinpublishedmapsandinstitutionalaffiliations.
Frontcoverpicture:PerformanceattheWaterPuppetTheatre,Hanoi,courtesyofJohannesP.Schlöder.
ThisSpringerimprintispublishedbytheregisteredcompanySpringerNatureSwitzerlandAG
Theregisteredcompanyaddressis:Gewerbestrasse11,6330Cham,Switzerland
Preface
Thisvolumecontainsaselectionofpapersreferringtolecturespresentedatthe7th
International Conference on High Performance Scientific Computing held at the
Vietnam Institute for Advanced Study in Mathematics (VIASM) in Hanoi, on
March 19–23, 2018. The conference has been organized by the Institute of
Mathematics of the Vietnam Academy of Science and Technology, the
Interdisciplinary Center for Scientific Computing (IWR) of Heidelberg University
and the Vietnam Institute for Advanced Study in Mathematics.
HighPerformanceScientificComputingisaninterdisciplinaryareathatcombines
many fields such as mathematics, computer science and scientific and engineering
applications. It is a key high-technology for competitiveness in industrialized
countries, as well as for speeding up development in emerging countries. High
performance scientific computing develops methods for computational simulation
andoptimizationforsystemsandprocesses.Inpracticalapplicationsinindustryand
commerce,scienceandengineering,ithelpstosaveresources,toavoidpollution,to
reducerisksandcosts,toimproveproductquality,toshortendevelopmenttimesor
simply tooperate systems better.
Theconferencehadabout300participantsfromcountriesallovertheworld.The
scientific program consisted of 210 talks, presented in 16 invited mini-symposia
and contributed sessions. Eight talks were invited plenary lectures given by Mihai
Anitescu (Argonne National Laboratory), Jose Antonio Carrillo (Imperial College
London, now University of Oxford), William Cook (University of Waterloo),
Ekaterina Kostina (Heidelberg University), Nils Henrik Risebro (University of
Oslo), Adelia Sequeira (University of Lisbon), Eitan Tadmor (University of
Maryland) and Fredi Tröltzsch (Technische Universität Berlin).
Topics were mathematical modeling, numerical simulation, methods for opti-
mization and control, parallel computing including computer architectures, algo-
rithms, tools and environments, software development, applications of scientific
computing in physics, mechanics, hydrology, chemistry, biology, medicine,
transport, logistics, location planning, communication, scheduling, industry, busi-
ness and finance.
v
vi Preface
The participants enjoyed not only an intensive scientific program and discus-
sions but also versatile social activities, like a snake dinner, a performance at the
water puppet theatre, a boat tour around Trang An at Ninh Binh and excursions to
theImperialCitadel ofThăngLongandtheVietnam MuseumofEthnology.Asat
the previous conferences, a satellite workshop on Scientific Computing for the
Cultural Heritage, including a field visit to the temples of Angkor, was jointly
organizedbytheRoyalUniversityofPhnomPenhandtheIWRattheConservation
d’Angkor Centre, Siem Reap, Cambodia.
The submitted manuscripts have been carefully reviewed and 18 of the contri-
butions have been selected for publication in these proceedings. We would like to
thank allcontributors andreferees.Our special thanks gotoNam-Dũng Hoangfor
his invaluable assistance in preparing this proceeding volume.
We would also like to use the opportunity to thank the sponsors whose support
significantly contributed to the success of the conference:
(cid:129) Berlin-Brandenburg Academy of Sciences and Humanities (BBAW),
(cid:129) CenterforModelingandSimulationintheBiosciences(BIOMS)ofHeidelberg
Universtity,
(cid:129) Faculty of Computer Science and Engineering of the Ho Chi Minh City
University of Technology,
(cid:129) Institute of Mathematics of the Vietnam Academy of Science and Technology,
(cid:129) Interdisciplinary Center for Scientific Computing (IWR) of Heidelberg
University,
(cid:129) International Council for Industrial and Applied Mathematics (ICIAM),
(cid:129) Vietnam Academy of Science and Technology (VAST),
(cid:129) Vietnam Institute for Advanced Study in Mathematics (VIASM),
(cid:129) Zentrum für Technomathematik (ZeTeM) of the University of Bremen.
Heidelberg, Germany Hans Georg Bock
June 2020 Willi Jäger
Ekaterina Kostina
Hoang Xuan Phu
Contents
GlobalOptimizationApproachfortheAscentProblemofMulti-stage
Launchers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
O. Bokanowski, E. Bourgeois, A. Désilles, and H. Zidani
A Robust Predictive Control Formulation for Heliogyro Blade
Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Adonis Pimienta-Penalver, John L. Crassidis, and Jer-Nan Juang
Piecewise Polynomial Taylor Expansions—The Generalization
of Faà di Bruno’s Formula. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Tom Streubel, Caren Tischendorf, and Andreas Griewank
Grid-Enhanced Polylithic Modeling and Solution Approaches
for Hard Optimization Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Josef Kallrath, Robert Blackburn, and Julius Näumann
Model Predictive Q-Learning (MPQ-L) for Bilinear Systems. . . . . . . . . 97
Minh Q. Phan and Seyed Mahdi B. Azad
SCOUT: Scheduling Core Utilization to Optimize the Performance
of Scientific Computing Applications on CPU/Coprocessor-Based
Cluster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Minh Thanh Chung, Kien Trung Pham, Manh-Thin Nguyen,
and Nam Thoai
Chainer-XP: A Flexible Framework for ANNs Run on the Intel®
Xeon PhiTM Coprocessor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Thanh-Dang Diep, Minh-Tri Nguyen, Nhu-Y Nguyen-Huynh,
Minh Thanh Chung, Manh-Thin Nguyen, Nguyen Quang-Hung,
and Nam Thoai
Inverse Problems in Designing New Structural Materials . . . . . . . . . . . 149
Daniel Otero Baguer, Iwona Piotrowska-Kurczewski, and Peter Maass
vii
viii Contents
Coupled Electromagnetic Field and Electric Circuit Simulation:
A Waveform Relaxation Benchmark . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Christian Strohm and Caren Tischendorf
SCIP-Jack: An Exact High Performance Solver for Steiner Tree
Problems in Graphs and Related Problems . . . . . . . . . . . . . . . . . . . . . . 201
Daniel Rehfeldt, Yuji Shinano, and Thorsten Koch
Physical Parameter Identification with Sensors or Actuators
Spanning Multiple DOF’s. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
Dong-Huei Tseng, Minh Q. Phan, and Richard W. Longman
Monotonization of a Family of Implicit Schemes for the Burgers
Equation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
Alexander Kurganov and Petr N. Vabishchevich
The Insensitivity of the Iterative Learning Control Inverse Problem
to Initial Run When Stabilized by a New Stable Inverse . . . . . . . . . . . . 257
Xiaoqiang Ji and Richard W. Longman
Strategy Optimization in Sports via Markov Decision Problems . . . . . . 277
Susanne Hoffmeister and Jörg Rambau
An Application of RASPEN to Discontinuous Galerkin Discretisation
for Richards’ Equation in Porous Media Flow. . . . . . . . . . . . . . . . . . . . 323
Peter Bastian and Chaiyod Kamthorncharoen
On the Development of Batch Stable Inverse Indirect Adaptive
Control of Systems with Unstable Discrete-Time Inverse . . . . . . . . . . . . 337
Bowen Wang and Richard W. Longman
AnImprovedConjugateGradientsMethodforQuasi-linearBayesian
Inverse Problems, Tested on an Example from Hydrogeology . . . . . . . . 357
Ole Klein
Idiosyncrasies of the Frequency Response of Discrete-Time
Equivalents of Continuous-Time System . . . . . . . . . . . . . . . . . . . . . . . . 387
Pitcha Prasitmeeboon and Richard W. Longman
Global Optimization Approach for the
Ascent Problem of Multi-stage Launchers
O.Bokanowski,E.Bourgeois,A.Désilles,andH.Zidani
Abstract Thispaperdealswithatrajectoryoptimizationproblemforathree-stage
launcher with the aim to minimize the consumption of propellant needed to steer
thelauncherfromtheEarthtotheGEO(geostationaryorbit).Hereweuseaglobal
optimizationprocedurebasedonHamilton-Jacobi-Bellmanapproachandconsidera
completemodelincludingthetransferfromaGTO(geostationarytransferorbit)to
theGEO.Thismodelleadstoanoptimalcontrolprobleminvolvingalsosomeopti-
mizationparametersthatappearintheflightphases.First,anadequateformulationof
thecontrolproblemisintroduced.Then,wediscusssometheoreticalresultsrelated
tothevaluefunctionandtothereconstructionofanoptimaltrajectory.Numerical
simulations are given to show the relevance of the global optimization approach.
This work has been undertaken in the frame of CNES Launchers’ Research and
Technologyprogram.
1 Introduction
This paper concerns the design of a trajectory optimization procedure for space
shuttlesofAriane5type,withtheaimofsteeringapayloadfromEarthtotheGEO
orbit.
B
O.Bokanowski( )
Lab.J.-L.Lions,UniversityParisDiderot,5rueThomasMann,75013Paris,France
e-mail:[email protected]
E.Bourgeois
CNESLauncherDirectorate,52rueJacquesHillairet,75012Paris,France
e-mail:[email protected]
A.Désilles·H.Zidani
UnitédeMathématiquesAppliquées(UMA),EnstaParisTech,
828BddesMaréchaux,91762PalaiseauCedex,France
e-mail:[email protected]
H.Zidani
e-mail:[email protected]
©SpringerNatureSwitzerlandAG2021 1
H.G.Bocketal.(eds.),Modeling,SimulationandOptimizationofComplex
ProcessesHPSC2018,https://doi.org/10.1007/978-3-030-55240-4_1
2 O.Bokanowskietal.
Trajectoryoptimizationforaerospacelaunchershasbeenextensivelystudiedin
theliterature,seeforinstance[7, 9, 16, 32, 40]andthereferencestherein.
ThepioneeringGoddard[24]problemisperhapsthesimplestmodel.Itconsists
inmaximizingthefinalaltitudeoftherocket,foraverticalflight,withagivenini-
tial propellant allocation. In one dimension this model is described by three state
variables: the altituder of the launcher, its velocity v and its mass m. The system
−→
issubmittedtotheaerodynamicforce(thedrag F )andiscontrolledviathethrust
−→ D
force F . Since this work, several studies were made on theoretical properties of
T
the optimal trajectories [16, 17, 27] and numerical methods allowing to calculate
these trajectories [8, 9, 16, 27, 32, 34, 35, 40], and in particular [18, 28, 30, 38]
fortheascentproblem.Overthelast5decades,severalnumericalapproacheshave
been proposed and analyzed leading to some efficient methods for computing the
optimal trajectories. These methods are essentially based on two approaches. The
firstone,called“discretize-and-optimize”,consistsofdiscretizingthecontrolprob-
lem in order to recast it into a finite dimensional problem that can be solved by
a numerical optimization solver. This approach is easily applicable and its imple-
mentationiswidelyaccessible,asmanyefficientoptimizationsolversareavailable
(Ipopt,WORHP,…).However,ingeneral,theoptimizationalgorithmsrequiresome
smoothnessoftheobjectivefunctionandsomequalificationconditionsonthecon-
straints.Moreover,theinitializationoftheiterativeprocessmaybechallengingfor
somecases,andthereisnoguaranteethatthecomputedsolutionisaglobalsolution,
unlesstheoptimizationproblemenjoyssomeconvexityproperties.
Another very used approach, the shooting method, consists in solving the opti-
malitysystemderivedfromthecontinuousproblem(i.e.,beforediscretization).This
optimalitysystemleadstoatwo-boundarydifferentialsystemthatmightbesolved
byaNewton-typemethod.Whiletheimplementationofthemethodisquitesimple,
itmayrequireanaprioriknowledgeofthecommandstructure(numberofswitching
times,existenceofsingulararcs,etc.…).Unlessforspecificcontrolproblems,the
structureofcontrollawsisachallengingquestion.
In this work, we investigate the resolution of the ascent problem by a third
approach:theHamilton-Jacobi-Bellman(HJB)approach.ItisbasedontheDynamic
Programming Principle (DPP) studied by R. Bellman [6]. It leads to a characteri-
zationofthevaluefunctionasasolutionofanHJBequationwhichisafirstorder
nonlinearpartialdifferentialequation(PDE)indimensiond,whered isthenumber
of state variables involved in the problem. The HJB equation may be viewed as a
differentialformoftheDPP.Animportantbreakthroughforthisapproachoccurred
inthe80’s,whenthenotionofviscositysolutionsofnonlinearPDEswasintroduced
byCrandallandLions[19–21].Thistheoryallowstoestablisharigorousframework
forthetheoreticalandnumericalstudytheHJBequationsarisinginoptimalcontrol
theory. The contributions in this direction do not cease growing, see the book of
BardiandCapuzzo-Dolcetta[5].
An interesting by-product of the HJB approach is the synthesis of the optimal
controlinfeedbackform.OncetheHJBequationissolved,foranystartingpoint,the
reconstructionoftheoptimaltrajectorycanbeperformedinrealtime.Furthermore,