Prof. Dr. Gabriela P. Henning
– Santa Fe
Gabriela Henning’s Biosketch
Gabriela P. Henning is a Professor and Researcher at INTEC
(CONICET-Universidad Nacional del Litoral), Santa Fe, Argentina, where she works
at the “Laboratory
for Intelligent Systems for the Optimal Design and Operation of Industrial
Processes”. She obtained a Chemical Engineering Degree from “Universidad
Tecnológica Nacional, Facultad Regional Rosario”, Rosario, Argentina, in
1981 and a Ph.D. in Chemical Engineering from “Universidad
Nacional del Litoral”, Santa Fe, Argentina, in 1986. She was a
Postdoctoral Fellow at the Laboratory for Intelligent Systems in Process
Engineering at the Massachusetts Institute of
Technology (1986-1989), specializing in the application of Artificial
Intelligence techniques in Process Systems Engineering where she and Dr.
Horacio P. Leone initiated the MODEL.LA
project under the supervision of Prof. George Stephanopoulos.
She is actively involved in teaching
activities at the Industrial Engineering Degree Program of Universidad Nacional
del Litoral, where she is the Head of that program. Her present teaching
interests include undergraduate courses in Supply Chain Management and
Information Systems for Manufacturing as well as a graduate course on
Information Systems for Operations Management.
The research interests of Prof.
Henning include the following topics:
Integrated information systems
for production environments. Languages for Enterprise Modeling.
Product Modeling within the
framework of industrial information systems.
Development of an environment
for modeling and managing the design process.
Scheduling: Application of
knowledge-based, mathematical programming and constraint-based methodologies to
Vehicle routing and
scheduling: Application of knowledge-based, mathematical programming and
These themes are addressed by
applying different methodologies and tools originated from the fields of
conceptual modeling, software engineering, artificial intelligence and
mathematical programming. Prof. Henning has authored or co-authored more than 30
papers, published in refereed journals or as book chapters, as well as more than
60 presentations in National and International Conferences. She has also been
involved in consulting activities, being responsible for the implementation of
various decision-support systems in the area of planning and scheduling.
information systems for production environments. Languages for Enterprise
Gabriela P. Henning
Horacio P. Leone
Gabriela S. Mannarino (Former Ph. D. Student)
strategies, such as quality improvement, business process reenginering and
enterprise integration are currently employed by production organizations
to cope with a highly competitive environment. Though conceptually different,
they share one common aspect: the need to understand and describe the target
organization though its objectives, process, resources, costs, etc. This
knowledge can be capture by developing different models of the organization.
are basic tools to analyze and evaluate the way activities are performed and to
improve and restructure them if needed. Models can be used by organizations as a
mechanism to acquire knowledge about the processes carried out at the production
floor as well as about the management functions that make possible these
production processes. The use of a language shortens the modeling process as it
defines the vocabulary to use and the way this vocabulary can be combined to
describe an organization.
Goals and Aims
address the needs mentioned above, this project proposes the development of a
language, named Coordinates, for describing a production organization through
its various dimensions. Within Coordinates, Task, Domain and Dynamic models
represent both the static and dynamic aspects of an industrial environment. Each
one gives a partial view, but all of them are required to get a complete outlook
of the organization. Domain models are used to identify the relevant entities
(i.e. distinct resources, services, organizational units) of a production
enterprise, to characterize them and to represent the static relationships among
them. In turn, Task models describe both the administrative and process
operation activities (tasks) of an organization in terms of a set of resources
that participate with specific roles (i.e., use, create, eliminate, modify,
etc.) and goals that tasks attempt to materialize. Tasks, which may be related
through temporal links, can be described at various abstraction levels,
according to the problem at hand. A Task can adopt several forms as (i)
alternative types of resources may be available, (ii) the same group of
resources may be combined differently, (iii) different tasks may be invoked to
achieve the same goals, etc. Thus, the TaskVersion concept is introduced in the
Coordinates language. Dynamic models focus on the behavior of a given resource
and on the interactions among a set of resources in a specific scenario.
Information System Modeling in Production Environments", G.S. Mannarino, G.P. Henning and H.P. Leone,
Infrastructure Systems for Manufacturing, J.B.M. Goosenaerts, F. Kimura and
H. Wortman (Eds), IFIP- Chapman & Hall, 103-114 (1997).
Information Systems: Modeling Support Tools”, G.S.
Mannarino, G.P. Henning and H.P. Leone. Computers and Chemical Engineering,
21, S667-S672 (1997).
“Multiview Enterprise Models
for Process Industries", G.S.
Mannarino, G.P. Henning and H.P. Leone. Latin American Applied Research, 28,
Coordinates: A Language for Enterprise Modeling",
G.S. Mannarino, G.P. Henning and H.P.
Leone, Information Infrastructure Systems for Manufacturing II, ,
J.J.Mills, F. Kimura (Eds), IFIP-Kluwer Academic Publishers, 379-390, (1999).
“A Task-Resource Based Framework for Process Operations
Modeling”, G..S. Mannarino, H.P. Leone
and G.P. Henning, Foundations of Computer Aided Process Operations 1999,
AIChE Symposium Series 320, Vol.94, AIChE- CACHE Corp. 279-285 (1999).
"Coordinates. Un Lenguaje para el Modelado de Empresas", Doctoral Dissertation in Computer Science, Computer Science Department, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, aproved with honors on December 18, 2001 by Gabriela Mannarino, under the supervision of Dr. Gabriela P. Henning (Advisor) and Dr. Miguel Felder (Co-advisor).
“Coordinates: A Language for
Enterprise Modeling”, G.S.
Mannarino, G.P. Henning and H.P. Leone. Proceedings of ICEIS 2001, Third
International Conference on Enterprise Information Systems, Setúbal,
Portugal, Vol. 2, 627-632 (2002).
Modeling within the framework of industrial information systems
Horacio P. Leone
Gabriela P. Henning
Marcela Vegetti (Ph. D. Student)
Motivation and challenges
Product proliferation is uncontrollable nowadays. Mass production of identical products is no longer viable for many industries. Products that were formerly quite standard, turn now into custom-made ones, leading to an enormous increase in product variants. In addition to the enormous variety of product types, products life cycle has been shrinking dramatically. These phenomena, which have been referred as product flexibility have become a new factor in competition. However, product flexibility has an enormous impact in supporting information systems, which are burdened by the growing number of products that must be managed. Indeed, high degrees of redundancy can occur in data management when closely related product structures are treated independently.
Market and technological forces affecting today’s competitive environment are changing dramatically. Information systems and computer technology have become basic tools to face these changes. Incorporation of this technology in production management has been evolving through successive generations of information systems: material requirements planning (MRP), manufacturing resource planning (MRP II), enterprise requirements planning (ERP), etc. An essential building block for the successful implementation of these systems, the bill of materials (BOM) is now more than a simple product structure. Actually, product data is used as the basis for many modules of an industrial information system.
Another problem faced by many industries in the fact that conventional BOM representations are only suitable for discrete manufacturing industries where products are always fabricated by putting parts together (composition) in assembly processes. In other words, conventional BOMs do not handle representations where products are obtained by decomposing raw materials, like in some food industries (milk and meat ones) and in the petrochemical business where hybrid structures (combining composition and decomposition types of operations) may be associated to products.
Goals and Aims
The problems previously described reveal a demand for a new representation of BOMs, able to suit the needs of different functional areas, to efficiently deal with a growing number of product variants and to handle all types of production strategies. Associated with such representation, there is need for the corresponding bill of materials processor, the specific computer software that deals with data entry, maintenance and recovery. These requirements are guiding the development of an object-oriented framework for Bills of Materials in Process Industries.
The proposed product model tries to represent all the legitimate company uses of product information managing crucial aspects that should be taken into account in a BOM representation, such as (i) efficient handling of variants and restrictions, (ii) ease with which new product structures are derived and (iii) possibility of representing decomposition BOMs (associated with production processes involving the decomposition of a raw material) and hybrid BOMs (combination of composition and decomposition processes). It is being implemented using OODBMS technology that allows the creation of persistent objects, enabling the implementation of the object model as it was conceived, without transforming it into a relational outline.
Object-Oriented Framework for Bills of Materials in Process Industries”, M. Vegetti, G.P. Henning and H. Leone,
Process Engineering-10, J. van Schijndel and J. Grievink (Eds), Elsevier,
An Object-Oriented Model for
Complex Bills of Materials in Process Industries”, M. Vegetti, G.P. Henning and H. Leone, Brazilian Journal of Chemical Engineering,
No. 4, 491-497 (2002).
of an environment for modeling and managing the design process
Gabriela P. Henning
Horacio P. Leone
Silvio Gonnet (Ph. D. Student)
Motivation and Challenges
Design problems (DP) are
inherently complex and ill defined. Therefore, the structure of the design
process is not known in advance, it starts with a small set of requirements that
include goals and constraints and evolves through subsequent stages of
increasing complexity in a non-linear manner. In most cases there is a lack of a
fully articulated methodology, so, there is no clear distinction between
During a design process (DPR),
models of the artifact being designed are generated. They differ in granularity,
complexity and associated assumptions; therefore, there is an explicit need to
properly manage model versions.
Unfortunately, once a design
project is finished, the things that remain are mainly "design
products" such as the models that were generated, detailed specifications
of the resulting artifact, drawings, sketches, etc. However, there is no
explicit representation of how they were obtained. More specifically, there is
no trace of which requirements were imposed, which activities originated a given
product, which actors performed a given activity, which is the underlying
rationale behind a decision-type of activity, etc.
Due to their size and
complexity or specific needs of expertise, DP are rarely tackled by individuals,
design teams are the usual coin. Thus, human experts along with computer-aided
tools are the ones that by interacting cooperatively, sharing resources of
various type and design products, solve complex problems. This poses the need
for handling the DPR as a cooperative one.
Goals and Aims
In order to tackle the problem
described above this project proposes a framework for representing and capturing
the design process. This framework would act as a foundation for developing
computational tools to support the DPR and to guide designers in the different
activities of a design project.
The proposed framework is
defined in terms of metamodels that allow the representation of the executed DPR
and the evolution of the different design objects that participated in it.
Design objects may be design products as well as the requirements that specified
them, or argumentations and goals posed by actors when they performed a given
activity. Metamodels can be specialized according to the particular domain being
tackled. It can be done in terms of the different operations that are applied to
the distinct design objects, and in terms of the different design objects that
participate in the DPR.
Situational calculus in conjunction with the object-oriented paradigm
let us model experts´ knowledge and their particular rationale in relation to a
given operation they applied. On the other hand, the extension of the IBIS model
allowed us to represent, at a higher level, the rationale behind a decision
taken during the DPR. Thus, these tools allow the tracing of the DPR and its
resulting products, as well as the analysis of the employed reasoning line,
setting the grounds for learning and future reuse.
In relation to this project,
cooperation has been established with the Process
Systems Engineering Group directed by Prof. Dr.-Ing. Wolfgang Marquardt
under the framework of the MODEPRO
"A Task and
Version-Oriented Framework for Modelling and Managing the Process Design
Process", S. Gonnet, G.S Mannarino, H.
P. Leone and G. P. Henning, AIChE Symposium Series 323, AIChE - CACHE,
"An Environment For Modeling And Managing The Process
S. Gonnet, G. S.
Mannarino, H. P. Leone and G. P. Henning, Latin American Applied Research,
Vol. 31, No. 5, 419-425 (2001).
“Modeling of Actors within a
Chemical Engineering Work Process Model”, M. Eggersmann, G.P. Henning, C. Krobb and H.P. Leone.
International CIRP Seminar, 6-8 June 2001, Stockholm, Sweden, 203-208.
“Representing and Capturing the Experts’ Knowledge in
a Design Process”, S. Gonnet, H. Leone and G.
Henning, Proceedings of ASAI’2002 (Argentine Symposium on
Artificial Intelligence), Santa Fe, Argentina, September
and Understanding Different Types of Process Design Activities”, M.
Eggersmann, S. Gonnet, G.P. Henning, C. Krobb,
and W. Marquardt, to appear in Latin American Applied Research,
Vol. 33, No 2, 167-175 (2003).
Application of knowledge-based, mathematical programming and constraint-based methodologies to industrial problems
Gabriela P. Henning
Luis Zeballos (Ph. D. Student)
Motivation and Challenges
Renewed interest in production
scheduling has been stimulated by many factors. The most critical ones are
enterprise attempts for optimizing their overall supply chains in response to
globalization and competition. Better production schedules provide a competitive
advantage through gains in resource productivity, improved efficiency in
operations management, and higher customer satisfaction.
Real-world scheduling problems
are inherently difficult because of the dynamic nature of industrial
environments, conflicting organizational goals, the existence of operational
constraints and preferences that are difficult to represent in a computational
model as well as the intrinsic complexity of any scheduling problem. Scheduling
includes both creating a schedule for a production facility (predictive
scheduling) and adapting an existing schedule when unforeseen events occur
The dynamic nature of the
scheduling problem has two different roots. One is the unpredictability of the
execution environment. In industrial plants, no matter how well defined a
predictive schedule is, reactive facilities are needed to cope with unexpected
events/disturbances on the shop floor as well as changes in production orders.
The other root comes from the intrinsic characteristics of the problem.
Scheduling is not an isolated, stand-alone function. Quite the opposite, it
involves multiple decision-makers that belong to different departments of the
organization and generally pursue distinct competitive goals. For them, a
schedule is a context for identifying conflicts or weak points of a plan and for
Goals and Aims
In order to tackle the
scheduling of industrial facilities several research lines have been adopted,
each one aiming at the following goals:
This research line presents a knowledge-based framework for scheduling systems aimed at solving a rather broad range of industrial problems. It employs an explicit representation of the domain and a hierarchical, task-oriented approach for the solution of predictive and reactive scheduling problems. The ideas behind the framework have been successfully applied in the implementation of three different support systems, two of which have already entered into every-day industrial use. The framework's architecture has three layers (i) domain knowledge layer, (ii) problem solving layer and, (iii) graphical user interface. It is important to remark that the same architecture supports both predictive and reactive scheduling activities. The reactive scheduling function is envisioned as an interactive schedule-repair task. This is possible because the outcome of the scheduling task is explicitly represented in the framework. Interactive facilities are aimed at supporting the construction and evolutionary modification of schedules by means of mouse-and-click actions over Gantt diagrams. The proposed approach pursues the engagement of the user in the process of maintaining a schedule in a dynamic environment, in a non-disruptive fashion, by providing him/her with interactive facilities for schedule modification. The proposed framework relies on an explicit object-oriented representation of the schedule and supports, up to now, three categories of user-driven revision actions: operation-based, production order-based and resource-based schedule modification actions. Before executing any action proposed by the scheduler possible conflicts are checked. A rich underlying representation of the domain layer keeps track of different kinds of soft and hard constraints and prevents users from possible mistakes.
Programming) Based Approach
Constarint-logic programming (CP) is a technique that has gained increased attention in the last years and that has been successfully applied to the scheduling domain. Constraints allow programmers and users to think in high level terms by stating declaratively the relations that should be maintained in a given problem. The possibility of using efficient constraint based languages and solvers, like ILOG Solver, having specially designed components, like ILOG scheduler, that provides specialized modeling and algorithmic enhancements for solving scheduling problems, is being tested in this project. Within this project new a CP formulation has been proposed to tackle real industrial problems. The synergic approach of adding to a CP formulation some constraints of mathematical programming type has been successfully tested in problems associated to multiproduct batch plants. Features commonly found in industrial facilities (e.g. sequence-dependent changeover times, finite equipment and orders release times, existence of limited renewable and non-renewable resources) are taken into account.
Mathematical Programming Based
This research line, which is directed by Prof. Jaime Cerdá, pursues the development of systematic methodologies for the scheduling of single-stage and multi-stage batch facilities. The proposed approach does not rely on the definition of time slots or events; instead it is based on a continuous time domain representation.
"A Knowledge-Based Approach to Production Scheduling for Batch
Henning y J. Cerdá. Computers Chem. Engng,
20, Suppl. B, S1295-S1300 (1996).
"An Expert System for the Short-Term Scheduling of Multistage
Multiproduct Plants Manufacturing Assorted Products". G. P. Henning y J. Cerdá ,
Symposium Series 312, Vol. 92, 397-400 (1996).
Mixed-Integer Linear Programming Model for Short-Term Batch Scheduling in
Parallel Lines". J. Cerdá, G.P. Henning and I.E. Grossmann.
Res, 36, No. 5, 1695-1707 (1997).
predictive and reactive scheduling in industrial environments",
Henning and J. Cerdá, Computers Chem. Engng., 24,
9-10, 2315-2338 (2000).
Scheduling of Batch Plants Satisfying Multiple Product Orders with Different
Continuous-Time Approach to Short-Term Scheduling of Resource Constrained
Multi-Stage Batch Facilities", C.A. Méndez, G.P. Henning y J. Cerdá,
Engineering-8, S. Pierucci S (Ed.), Elsevier Science Ltd., 1045-1050 (2000).
Interactive Scheduling of Multiproduct Batch Plants”, G.P.
Henning, Lecture Notes in Artificial Intelligence 1952, M. C.
Monard, J. S. Sichman (Eds), Springer-Verlag, 76-85 (2000).
of Interactive Facilities in a Knowledge-Based Scheduling Framework”,
G. P. Henning, Computer Aided Process
Engineering-11, S. Jorgensen and R. Gani
(Eds), 883-888 (2001).
MILP Continuous-Time Approach to Short-Term Scheduling of Resource-Constrained
Multistage Flowshop Batch Facilities”. C.A.
Méndez, G.P. Henning and J. Cerdá. Computers Chem. Engng , Vol. 25,
4-6, 701-711, (2001).
Scheduling of Multiproduct Batch Plants Under Limited Resource Capacity”.
Méndez, G.P. Henning and J. Cerdá. Latin American Applied Research Vol. 31,
No. 5, 455-462 (2001).
"A Constraint Programming Approach to the Single-Stage Scheduling Problem with Resource Constraints", L. J. Zeballos and G.P. Henning, presented in SIO (Simposio de Investigación Operativa) held during 31 JAIIO (Jornadas Argentinas de Informática e Investigación Operativa), Santa Fe, Argentina, September 2002.
"A Constraint Programming Approach to the Multi-Stage Batch Scheduling Problem", L.J. Zeballos and G.P. Henning, Proceedings of FOCAPO 2003, Coral Springs (Florida, USA) January 12-15, 2003.
routing and scheduling: Application of knowledge-based, mathematical programming
and constraint-based methodologies.
Gabriela P. Henning
Páez de la Torre (Industrial Engineering Undergraduate Student)
González Rossia (Industrial Engineering Undergraduate Student)
Motivation and Challenges
in better approaches to vehicle routing and scheduling (R&S)
been stimulated by a variety of factors. Perhaps, the most important one is
globalization and its direct impact in every logistic function. Nowadays,
logistic costs account for 12 % of product costs. Only in Europe there are over
12000 logistic companies having an annual turnover of 1.200 billion EU. To
account for the magnitude of the business consider that in 1990 13 million
trucks where responsible of a movement of 800 billion ton-kilometers around
Europe. Better solutions to R&S problems provide a competitive advantage
through reductions in costs, associated efficiencies in operations management as
well as customer satisfaction..
Competition has also motivated the development of more efficient, flexible and
integrated computer support tools. Indeed, there is a real need for better
support environments, whose merit must be measured in terms of a variety of
features such as power (solution quality versus response time), configurability,
adaptability to a variety of problems, integration with other systems, map
visualization capabilities, user interaction and friendliness, reactivity,
possibility of introducing manual changes, etc.
Real world R&S problems are inherently complex. They need to address a wide variety of constraints and preferences that may be difficult to express in a computational model and are associated to a set of sometimes conflicting performance measures. Therefore, rigid R&S solution procedures, oriented towards the achievement of optimal or sub-optimal solutions, may not always be adequate. Moreover, purely automatic R&S procedures may not be realistic enough since they may neglect the crucial role of a human expert, who has the responsibility for all decisions and likes to be engaged in the solution process.
Goals and Aims
the difficulties pointed out before, this project aims at developing a knowledge-based environment that is based on the object-oriented technology and
incorporates a variety of heuristic solution procedures and techniques.
system’s goal is to provide a friendly setting where the user can (i) easily
define and characterize a variety of R&S problems by making appropriate
selections in a menu-oriented interface, (ii) resort to a variety of solution
methodologies to generate and explore alternative routes for the problem being
studied, (iii) assess the quality of each obtained solution by means of
different performance measures and a visual representation, (iv) ameliorate the
quality of a given solution by means of improvement operators or (v) manually
modify a certain solution by means of user-driven actions.
The system’s goal is to provide a friendly setting where the user can (i) easily define and characterize a variety of R&S problems by making appropriate selections in a menu-oriented interface, (ii) resort to a variety of solution methodologies to generate and explore alternative routes for the problem being studied, (iii) assess the quality of each obtained solution by means of different performance measures and a visual representation, (iv) ameliorate the quality of a given solution by means of improvement operators or (v) manually modify a certain solution by means of user-driven actions.
Current and future tasks include the incorporation of better clustering and solution algorithms, the automatic generation and execution of mathematical models in a user-transparent way, the improvement of the existing help facilities, the automatic generation of sites’ adjacency information in order to improve the current implementation of intra and inter-route improvement operators as well as the definition of interfaces with commercial Geographic Information Systems.
"A user-friendly environment for routing and scheduling of vehicles", B. Páez de la Torre and G.P. Henning, Presented in SIO 2002 (Simposio de Investigación Operativa) held during 31 JAIIO (Jornadas Argentinas de Informática e Investigación Operativa), Santa Fe, September 2002.