Challenges in Real-time Optimization of Maintenance Decisions of Distributed Parallel Machines under Sustainable Production Constraints
Prof. Kondo Hloindo ADJALLAH
ENIM, LCOMS, FRANCE
The sustainable production approach should allow to increasing product quality while reducing the production costs through intelligent maintenance strategies enabling the zero-downtimes and low wastes in production machines, including low loss of energy. Indeed, very often, machines operating with degraded components increase the energy consumption and reduce the quality of products. Also, the increase of corrective maintenance frequency also increases failure hazards related to human factors. To allow machines to reach high availability and effectiveness in sustainable production units, the maintenance actions must consider a minimum of preventive maintenance actions and moreover of predictive maintenance, when the ratio of the cost over effectiveness is low. Among many others, this ratio is a useful feature to develop efficient and intelligent maintenance decision support algorithms for distributed parallel machines in sustainable production contexts. This keynote will consider advanced challenges in real-time optimization of maintenance decisions of distributed parallel machines under sustainable production constraints. Specific algorithms and some heuristics will be suggested for facing these challenges, while taking advantage from new features extracted from machines operation and environment data, and data of maintenance activities. In the last step of this keynote, maintenance logistics variables will be considered in the decision support model.
Kondo H. Adjallah received the M.S. degree in electrical and control engineering. He earned the Ph.D. with the National Polytechnic Institute of Lorraine, France. He has been Associate Professor at the Université de Technologie de Troyes, France, from 1994 to 2008. Since September 2008, he got the full Professor position at the Ecole Nationale d’Ingénieurs de Metz, where he is leading research activities at the LCOMS Laboratory of the University of Lorraine. His is visiting professor at University of Cincinnati. He contributed successively to research activities with the CRAN of Nancy, Charles Delaunay Institute of Troyes and LGIPM at the University of Metz. His current research interests include data integration and analysis, reliability and degradation modeling, health condition monitoring and decision support for predictive maintenance. He is leading currently research activities on data driven modeling of risks, anticipation and resiliency, with focus on industrial data instrumentation, data collection, health information extraction, dependability decision support of sustainable infrastructures. He has been chairman of the French Research Working Group on “modeling and optimization of the distributed & collaborative maintenance” of the CNRS affiliated research group on “modeling, analysis and steering of dynamic systems” (GRD MACS). He is member of the IEEE society since 2002 and leader of the international cooperative research network for the management of durable infrastructures of development since 2007. Also, he is fellow of the NSF Industry/University Cooperative Research Center on Intelligent Maintenance Systems.
Big Data Analytics for Digital Decision and Smart Production
Professor Chen-Fu CHIEN
Department of Industrial Engineering & Engineering Management,
National Tsing Hua University, Hsinchu 30013, TAIWAN
The increasing adoption of the multimode sensors,intelligent equipment and robotics, mobile communication devices, “Internet of Things” (IOT) and big data analytics have enabled an unprecedented level of manufacturing intelligence for smart production. The introduction of new IOT and information technologies and the advances in big data analytics capabilities are having profound effects on the decisions, production and service operations involved in digital manufacturing. Indeed, various companies are battling for dominant positions in this newly created arena via providing novel solutions and/or employing new technologies to enhance digital manufacturing. On the basis of our extensive collaborative studies with high-tech industries, this keynote aims to address emergent issues driven by the needs of big data analytics for digital decision and smart production in advanced intelligent manufacturing systems such as yield enhancement, defect diagnosis, quality control, advanced equipment control, cycle time reduction, cost reduction, human capital and productivity enhancement with empirical studies for illustrations. This talk will conclude with discussions of the implications of evolutionary digital manufacturing technologies and applications to foster more discussions of shifting paradigms for computers and industrial engineering.
Dr. Chen-Fu Chien is Tsinghua Chair Professor, in the Department of Industrial Engineering & Engineering Management, National Tsing Hua University (NTHU), Taiwan. Professor Chien is the Principal Investigator for the Semiconductor Technologies Empowerment Partners (STEP) Consortium sponsored by Ministry of Science & Technology and the Director for the NTHU-TSMC Center for Manufacturing Excellence. He received B.S. with double majors in Industrial Engineering and Electrical Engineering with the Phi Tao Phi Honor from NTHU in 1990. He received M.S. in Industrial Engineering and Ph.D. of Decision Sciences and Operations Research with two minors in Statistics and Business at the University of Wisconsin-Madison, U.S.A., in 1994 and 1996, respectively. He was a Fulbright Scholar in the Department of Industrial Engineering and Operations Research, UC Berkeley, from 2002 to 2003. He also received the Executive Training of PCMPCL from Harvard Business School in 2007. He joined NTHU as Assistant Professor in 1996, was an Associate Professor from 1999 to 2003, and has become a Full Professor since 2003. Then, he was promoted as a Tsinghua Distinguished Professor from 2010 to 2013, and has become Tsinghua Chair Professor since 2013. He was a Visiting Professor in Institute for Manufacturing, Cambridge University (sponsored by Royal Society, UK), Visiting Professor in Beijing Tsinghua University (sponsored by Chinese Development Foundation), Visiting Professor in Waseda University (sponsored by Japan Interchange Association Young Scholar Fellowship), and Visiting Adjunct Professor in Tianjin University and Zhejiang University, China.
His research mainly concerns the development of decision analytics and optimization solutions for companies confronting with decision problems involved in strategy, manufacturing, and technology that are characterized by uncertain (incomplete information or Big data) and a need for tradeoff among various objectives and the information system for effective implementation. Dr. Chien and his Decision Analysis Lab Associates have been actively involved in many university-industry collaborative research projects in the high-tech industries including semiconductor manufacturing, solar, LED, and computers in which they employ their expertise in solving real problems with domain experts. From 2005 to 2008, he had been on-leave to serve as the Deputy Director of Industrial Engineering Division in Taiwan Semiconductor Manufacturing Company (TSMC) that is the world leading wafer foundry. Dr. Chien has applied several invention patents for semiconductor manufacturing methodologies, eight of which have been granted. Dr. Chien has received many awards including the National Quality Award from the Executive Yuan (2012), twice Distinguished Research Awards (2007, 2011), Tier-One Principal Investigator (2005-2008), and Best Research Awards from the National Science Council, University Industrial Contribution Award from Ministry of Economic Affairs for Individual Contribution (2009) and Group Contribution (2010), Distinguished University-Industry Collaborative Research Award from the Ministry of Education (2001), Distinguished Young Faculty Research Award (2001) and Distinguished University-Industry Collaborative Research Award (2007) by NTHU, Best Paper Award (2001), Distinguished Young Industrial Engineer Award (2001), IE Medal (2010) from Chinese Institute of Industrial Engineers, Best Engineering Paper Award (2002) and Distinguished Engineering Professor (2010) by Chinese Institute of Engineers, TSMC-NTHU Faculty Semiconductor Research Grant (2004), and the Lu, Feng-Chang Award from Chinese Management Association (2007). Dr. Chien is a member of IEEE, CIIE, SEED, CIDS, and CSMOT and is a Board Member of CIIE, SEED, and CIDS. Dr. Chien has served as the Steering Committee of the Industrial Engineering and Management Program in National Science Council since 2002. He is Area Editor for Flexible Services and Manufacturing Journal, Editor for Computers and Industrial Engineering and Associate Editor for IEEE Transactions on Automation Science and Engineering and Journal Intelligent Manufacturing. He is also on the Advisory Board of OR Spectrum and editorial board for International Journal of Operational Research (UK), Journal of the Chinese Institute of Industrial Engineers (EI/TSSCI), Sun Yat-Sen Management Review (TSSCI), Journal of Management and Systems (TSSCI), and Journal of Quality (EI).
Applications of Metaheuristics to Manufacturing
Scheduling for HDD, Panel and Semiconductor Devices
Fuzzy Logic Systems Institute and Tokyo University of Science, JAPAN
Many Combinatorial Optimization Problems (COP) in the real world manufacturing systems impose on more complex issues, such as complex structure, nonlinear constraints, and multiple objectives to be handled simultaneously and make the problem intractable to the traditional approaches because of NP-hard COP. In order to develop an efficient solution algorithm that is in a sense "good," that is, whose computational time is small, or at least reasonable for NP-hard combinatorial problems met in practice, we have to consider the following very important issues:
- Quality of solution,
- Computational time and
- Effectiveness of the nondominated solutions for multiobjective optimization problem (MOP).
Evolutionary Algorithm (EA) is a subset of metaheuristics, a generic population-based metaheuristic such as Genetic Algorithm (GA), Hybridized GA (HGA), Particle Swarm Optimization (PSO), and Estimation of Distribution Algorithm (EDA). EA is based on principles from evolution theory, and it is very powerful and broadly applicable stochastic search and combinatorial optimization technique which is effective for solving various NP hard COP models.
This talk will be firstly introduced a brief survey of Metaheuristics based on EA such as GA, HGA, Multiobjective GA (MoGA), PSO and EDA for applying various combinatorial optimization problems in real world manufacturing systems. Secondly real applications based on Metaheuristics will summarize the following recent manufacturing scheduling topics:
HGA and PSO+GA for Reentrant Flowshop Scheduling in HDD Manufacturing Scheduling,
Mo-HGA with TOPSIS for TFT-LCD Module Assembly Scheduling,
Cooperative EDA for Semiconductor Final Testing Scheduling, and
Hybrid EDA with multi-subpopulation for Semiconductor Manufacturing Scheduling.
Dr. Mitsuo Gen is a senior research scientist at Fuzzy Logic Systems Institute (FLSI) and visiting professor at Research Institute for Science and Technology, Tokyo University of Science (TUS), Japan. PhD in Electric Engineering, Kogakuin University, Japan in 1975 and PhD in Informatics, Kyoto University, Japan in 2006. Faculty: Ashikaga Institute of Technology 1975-2003 and Waseda University 2003- 2009. Visiting Faculty: University of California at Berkeley 1999-2000, Texas A&M University 2000, Hanyang University 2010-2012 and National Tsing Hua University 2012-2014, Tokyo University of Science 2014-current and Fuzzy Logic Systems Institute 2010-current. Research Field: Evolutionary Computation, Manufacturing Scheduling and Logistics. Society: Editor-in-Chief: IE & MS Journal, President of IMS: Inter. Assoc. of Info. & Mgmt. Sci. Area Editor:Computers & Industrial Engineering, Assoc. Editor: J. of Intelligent Manuf. Google Scholar Citation:Mitsuo Gen.
Partition Inequalities and Network Design
Prof. A. Ridha MAHJOUB
Laboratoire LAMSADE, FRANCE
Given a graph G=(V,E) with non-negative weights x(e) for each edge e,
a partition inequality is of the form x(δ(V1,…,Vp)) ≥ ap+b. Here δ(V1,…,Vp) denotes the multicut defined by a partition V1,…,Vp of V. Partition inequalities arise as valid inequalities for optimization problems such as survivable network design problems, and play a central role in solving these problems using cutting planes. In this talk, we attempt to survey some variants of these inequalities, examine different separation algorithms, and discuss applications in network design.
A. Ridha Mahjoub is Professeur Classe Exceptionnelle of Operations
Research and Combinatorial Optimization at Université Paris-Dauphine,
Paris, France. He is also member of the LAMSADE laboratory, CNRS.
Previous positions include full professor at the University of Brest,
France, from 1991 to 1998 and the University of Clermont-Ferrand,
France, from 1998 to 2007. Professor Mahjoub holds an undergraduate
degree in Mathematics from University of Tunis, Tunisia and a Ph.D.
and a Doctorat d’Etat in Operations Research and Combinatorial
Optimization from the University of Grenoble, France. His research
areas include the theory and applications of polyhedral approaches
for modelling, analysing and solving large NP-hard combinatorial
optimization problems, mixed integer programming as well as
complexity and graph theory. A part of his research has recently
focused on the design of cutting plane algorithms for network design
problems. Professor Mahjoub is author and co-author of several papers
that have appeared in leading journals such as Mathematical Programming,
Mathematics of Operations Research, SIAM Journal on Discrete Mathematics,
Discrete Mathematics, Discrete Applied Mathematics and Networks.
He served as co-director of the Mathematics and Computer Science
Department at Université Paris-Dauphine between 2008 and 2013.
Dr Mahjoub edited and co-edited books and several special issues of
journals. He currently serves as Editor-in-Chief of the international
journal RAIRO-Operations Research.
Operations Research : Old Problems, New Paradigms
Prof. Alain QUILLIOT
Laboratoire d’Informatique, de Modélisation et d’Optimisation des Systèmes
, Clermont-Ferrand, FRANCE
Operations Research was born at the end of the years 40’s, from the needs of American
Army supply chain, and deeply impregnated of the impact of mainframe technologies,
newly emerging linear programming methodology and centralized management. Time flew
away, and decisional computing has currently to adapt itself to Internet, wireless
communications and remote tracking technologies, to outsourcing, globalization and
highly distributed economics, as well as to a need for democratization and flexible
organization. This makes that very standard problems like for instance those which
are related to the management of mobility or the scheduling of industrial activities
have to be revisited while taking into account the change contexts which put focus on
the need on flexible and reactive answer, or on collaborative context. So the purpose
of this presentation is to introduce, through simple examples, the way new paradigms
related to robustness, mixed decision, collaborative or on line programming, may lead
to reformulate old problems into new ones, which open new challenges to searchers
while trying to get closer to industrial needs.
Alain Quilliot, born on the 25/10/1953, former student of ENS
Saint-Cloud (1973-77), « Agrégé de Mathématiques », and PhD graduate from PARIS VI
University (1978, Dir. C.BERGE), Professor Alain QUILLIOT started its academic career
at UNAM University of MEXICO (179-81) as Cooperating Professor, and next (1981-82),
at South Carolina University in COLUMBIA (USA) as Invited Professor. He next got a
permanent position at ENSIMAG GRENOBLE and, at the same time, defended his Thèse d’Etat
in June 1983 in PARIS VI. After he got in 1987 a Full Professor position at Blaise
Pascal University in CLERMONT-FERRAND, he created a Master/Engineering department in
Mathematics and Modelling inside the CUST/POLYTECH Institute, which he managed from
1989 until 1994. In 1993, he simultaneously created the Engineering Institute ISIMA,
member of the Concours Communs Polytechniques (ex ENSI, CPGE) Network,
and the inter-university research laboratory LIMOS (Informatics,
Modelling and System Optimization:
currently 85 permanent searchers, 94 PhD, 12 associated members and 7 technical staff),
which is a mixed CNRS/University Laboratory and which got A+ AERES evaluation in 2006 and
A evaluation in 2011. He led ISIMA from 1993 until 2007, and LIMOS from 1994 until 2014.
He also was, in 2003, a founding member of the TIMS CNRS Federation (Technologies of
Information, Mobility and Safety), which was made of 6 research groups
in Computing, Mechanics,
Robotics, which he led from 2007 until the end of 2011, and which gave rise
to the newly labeled
LABEX structure IMOBS3 (Intelligent and Innovative Mobility).
He is currently the head of the
CNRS network GDR 3002: Operations Research, which gathers
500 permanent searchers. Deeply
involved in international and industrial partnerships,
Professor QUILLIOT is also currently Chairman
of the International Affairs Office (M.R.I) of CLERMONT-FERRAND town,
Director of the CNRS French-Chinese
Joint Laboratory (LIA) Smart Computing for Sustainable Growth, and vice-chairman
cluster, which gathers regional I.T companies and academic players.
Its personal contributions as
a scientist include the supervision of 26 PhD and 9 HDR
(Research Supervision Agreement) thesis,
the redaction of more than 200 publications (75 in international
or national journals and 6 book
chapters), and several direct or indirect partnership agreements
with companies like EDF, THALES,
MICHELIN, France Telecom, SNCF, …
What if Inductive Data-Driven Algorithms Become IT Commonplaces?
About some epistemological, ethical and legal consequences
The classical epistemology of Information Technology —including Artificial Intelligence approaches— was traditionally based on deductive knowledge modelling and simulation, so that explanation and theory were never far from engineering. Things turn different within Inductive Data-Driven Algorithms, because they do not need anymore domain theories and explanation-based processes to succeed in providing useful results.
We claim that those new features introduce some important epistemological break in the IT classical environment, with important ethical and legal consequences.
Around some recent use-cases, we shall demonstrate how crucial it is to figure out those paradigmatic evolutions.
Dr. Francis Rousseaux is full professor at Reims Champagne-Ardenne University, France, and Visiting professor at Goldsmiths University of London. He is teaching artificial intelligence, decision support and knowledge management. As far as research is concerned, he is currently managing project in contemporary music & dance, territorial intelligence and defence. Before becoming a full professor, he was involved in IT business through multinational companies like Thales and EADS.