81. Andronico, A., L. Cossa, M. Gagliardi and C. Spera, "A Structured Model to Manage a Large Number of Transactions," European Journal of Operational Research, 86:3 (11/95), pp. 402-421. Presented at the Third International Workshop on Project Management and Scheduling (Como, Italy, July 6-8, 1992). Available as a technical report, Dept. of Mathematics, University of Siena (Italy), 20 pages, 1992.
This article applies the BLOOMS language (the subject of a Gagliardi-Spera item in this Bibliography) to the problem of managing the daily transactions arriving at the main computer center of an important Italian banking group.
82. Andronico, A., L. Cossa, M. Gagliardi and C. Spera, "An Object Oriented Approach to a Model Management System: Characteristics and Examples," in G. Bottaro and Z. Zoppoli (eds.), Proceedings of the 36th Annual ANIPLA Conference (Associazione Nazionale Italiana Per L' Automazione; Genoa, Italy, 16-18 November 1992). Available as a technical report, Dept. of Mathematics, University of Siena (Italy), 22 pages, 1992.
This paper rationalizes the transition from the authors' earlier model definition language ALM to their new object-oriented language BLOOMS. Their model management system MODASS also evolves to become object-oriented. The SM framework is retained. The appendix contains a discussion of model integration in BLOOMS using an example from item 8 of this Bibliography.
83. Andronico, A., P. Lomagistro and C. Spera, "Toward a Structured Modeling System," Working Paper 235, Dept. of Mathematics, University of Siena (via S. Bandini 25, 53100 Siena, Italy), 28 pages, 7/91. Presented at EURO XI (Aachen, Germany, 16-19 July 1991). See also P. Lomagistro and C. Spera, "ALM: un Linguaggio di Modellizzazione Basato sullo Structured Modeling," Atti del XXX Congresso Annuale AICA (Siena, Italy, 25-27 September 1991).
The authors have partially implemented a model management system called MODASS (MOdel Definition And Solution System) based on SM principles. It uses ALM, a variant of SML, as its model description language. One of the novelties of this paper is its focus on reducing model schemas to a kind of canonical form that is supposed to facilitate the automatic identification of schemas so that solvers can be invoked without need for manual selection by the user.
The Atti del XXX Cong. Annuale AICA paper contains details on ALM's grammar.
[Update 9/95: The software is maturing, and is in the process of being made accessible via the Internet. Gagliardi and Spera gave a presentation titled "MODASS: The Development Lines of a Model Management System" at the AILOG-AIRO 94 Annual Conference, Savona (Italy), 1994. In addition, they gave a presentation titled "The Features of a Model Management System" at the "Structured Modeling Implementations" session at the INFORMS International Singapore, June 1995.]
83A. Balasubramanian, P.R. and M.L. Lenard, “Structured Modeling Knowledge for Collaborative Environments,” Proceedings of the Thirty-First Hawaii International Conference on System Sciences, Vol. IV, IEEE Computer Society Press, Los Alamitos, CA, pp. 464-475, 1/98.
This paper concerns models (especially spreadsheet models) and related knowledge in the context of an organizational intranet where sharing, re-use, and collaboration are important. Structured modeling is proposed as a framework for organizing the content knowledge of models to be implemented as spreadsheets. There is a discussion of how to implement such an approach, and the benefits of doing so.
84. Bradley, G.H. and R.D. Clemence, Jr., "A Type Calculus for Executable Modeling Languages," IMA J. of Math. in Management, 1:4 (1987), pp. 277-291.
Typing and units of measurement are introduced into a simple SML-like modeling language in a rigorous, elegant, and active way. In return for doing some extra specification work, the model designer receives powerful consistency checks and useful services for automatic units conversion and scale factoring. The development is quite compatible not only with SM, but also with almost any algebraic modeling language. This work also illuminates the nature of model classes and model integration (see the next paper). Clemence has since completed his doctoral dissertation on this topic.
85. Bradley, G.H. and R.D. Clemence, Jr., "Model Integration with a Typed Executable Modeling Language," Proceedings of the Twenty-First Annual Hawaii International Conference on System Sciences, Vol. III, IEEE Computer Society Press, Los Alamitos, CA, pp. 403-410, 1/88.
This sequel to the previous paper applies the Bradley-Clemence type calculus to an important problem in model design and management: integrating separate but related, previously validated models. First it is shown how the type calculus can help to mechanize one of the burdens that falls to a modeler when composing a monolithic integrated model, namely preserving dimensional consistency. For this a simple SML-like modeling language is used. Then a simple, formal, integration specification language is sketched that avoids having to compose a monolithic integrated model. The key "library unit" concept is quite close to the concept of a "module" in SM, so applying the results of this paper in the SM context may be straightforward.
86. Chari, K. and T. Sen, "A Graph-Based Modeling System for Structured Modeling," working paper, 44 pages, 6/95. Revised 9/96. Not for publication (but see the next item of this Bibliography). Authors respectively at CIS/OM Area, College of Business, James Madison University, Harrisonburg, VA 22807 and at Dept. of Accounting, Pamplin College of Business, Virginia Tech, Blacksburg, VA.
This paper proposes a graph-based language and modeling system for structured modeling whose basic construct is the model graph for representing a model class. Model graphs incorporate a multi-level architecture that can represent models at varying levels of detail, allowing a completely graphical approach to representation and manipulation. At one level of detail, a model graph is very similar to a SML genus graph. At a higher level of detail, functions and tests are expressed using elementary functional and logical operators. These operator nodes enable the generic rules and functions of SML to be diagrammed rather than remain internal to function and test nodes. This facilitates syntax-directed graph editing. Model graphs can be defined recursively using module nodes, allowing a forest-like structure for models. This approach allows models to be built graphically either top-down or bottom-up.
Model graphs have three distinct types of edges to visually distinguish relationships between nodes. Some edges exhibit index inheritance, some value inheritance, and some simply show that two nodes are related (i.e., no index or value inheritance takes place).
To produce a model instance, a model graph can be instantiated by using data stored in a relational database to produce SML-like elemental detail tables that are 1:1 with model graph nodes.
Model graphs are formalized using an approach adapted from C. Jones' 2-part article in ORSA J. on Computing. The prototype implementation is the subject of a companion paper (the next item of this Bibliography).
87. Chari, K. and T. Sen, "GBMS/SM: A Graphical Modeling
Environment," working paper, 38 pages, 6/95. Revised 6/96.
Authors respectively at CIS/OM Area, College of Business, James Madison
University, Harrisonburg, VA 22807 and at Dept. of Accounting, Pamplin
College of Business, Virginia Tech, Blacksburg, VA. Presented at
the "Structured Modeling Implementations" session at the INFORMS International
Singapore, 6/95. This implementation paper, which is not for publication,
is -- with its companion paper, the previous item of this Bibliography
-- the basis for three published papers:
"A Graphical Modeling System: Applications in Organizational Model Management," Omega, 25:2 (1997), pp. 241-253; "An Integrated Modeling System for Structured Modeling Using Model Graphs," Journal of Computing, 9:4 (Fall 1997), pp. 397-416; and "An Implementation of a Graph-Based Modeling System for Structured Modeling (GBMS/SM)," Decision Support Systems, 22:2 (1998), pp. 103-120.
There is a 102-page "GBMS/SM v1.0: User Manual & Tutorials" dated 6/95 that thoroughly documents GBMS/SM. The software is available for research purposes.
88. Chari, S. and R. Krishnan, "Towards a Logical Reconstruction of Structured Modeling," Decision Support Systems, 10:3 (October 1993), pp. 301-317. An earlier version appeared in Proceedings of the Twenty-Third Annual Hawaii International Conference on System Sciences, Vol. III, IEEE Computer Society Press, Los Alamitos, CA, pp. 524-533, 1/90.
Recall that the core SM concepts admit many possible languages to represent them; SML is only one such language. This paper develops a fundamentally different executable language called LSM based on logic. One of the nice features of LSM is that, by treating SM elements and structures as objects, it can predicate information about models as well as express the models themselves (e.g., information concerning model history, assumptions, dimensional units, revision, integration, and validation). Additionally, LSM is embedded within a generic model management language called L-up introduced by Bhargava and Kimbrough. A partial Prolog implementation exists.
89. Davis, J., E. Gardner, S. Levy, M. Thomas, E. Subrahmanian and A. Westerberg, "Modeling Languages for Geoffrion's Structured Modeling," working paper, Engineering Design Research Center, Carnegie Mellon University, 10 pages, 8/95.
This paper applies POSTGRES and a system called n-dim to SM. n-dim (n-dimensional information modeling), originally designed and implemented as an environment for modeling and collaborative design work in such fields as engineering design, is based on three fundamental mechanisms: a flat space of objects, the ability to build modeling languages for all objects, and a concept of persistence called publishing. A notion of model acts as the primary information-structuring primitive. It turns out that the objectives of SM and n-dim have much in common, and that n-dim can serve as an SM implementation vehicle to provide for an interactive graphical modeling environment, the creation of well-formed SM instances, collaborative model sharing, and facilities for managing model revisions.
90. Davis, J., A. Srinivasan and D. Sundaram, "Support for Decision Making: The Design and Implementation of Modeling Systems," working paper, Dept. of Management Science and Information Systems, University of Auckland, New Zealand, 12 pages, 8/95.
This is a progress report on a project to implement SM using the object-relational DBMS Illustra, the commercial version of POSTGRES. The implementation places great emphasis on modularity, and provides a common predicate-calculus-style language for accessing both data and models. See the papers by Srinivasan and Sundaram later in this Bibliography.
An earlier progress report from the POSTGRES phase of the project contains different examples and places more emphasis on model integration: J. Davis, A. Srinivasan, D. Sundaram and B. Thompson, "Integrated Modeling Implementation Using Object-Relational Database Systems," Proceedings of the Workshop on Information Technologies and Systems '94, Vancouver, B.C.
91. Deng, D.D., "Structured Modeling: A Critique," Dept. of Business Studies, University of Buckingham, England, 14 pages, 1/91. Presented at APMOD91 (Applied Mathematical Programming and Modeling), London, 14-16 January 1991.
Unfortunately, the author was innocent of all work on SML and FW/SM, as well as of most of the considerable literature that has grown up around SM in recent years.
92. Desai, S., "Models as Data in an Extensible Database," Informal Note, John E. Anderson Graduate School of Management, UCLA, 109 pages, 5/22/88.
This paper, a dissertation proposal prelude, explores how the new technology of extensible database systems can be used as a platform for building a modeling environment. Examined in detail are three alternative designs for representing structured models: designs proposed by D. Dolk, A. Geoffrion, and M. Lenard.
93. Desai, S., "Are Extensible Databases Better Than Relational Database Systems for Model Management?", John E. Anderson Graduate School of Management, UCLA, 31 pages, 3/91.
Based on considerations of semantic power, manipulation tools, and storage efficiency, this paper argues that the answer to the question posed in the title is YES. Moreover, the advantages pertain not only to SM but also to other model representational styles such as AMPL, which is considered in some detail along with SM. This paper contributes to the author's thesis (see Part IIA of this Bibliography).
94. Dolk, D.R., "Data as Models: An Approach to Implementing Model Management," Decision Support Systems, 2:1 (March, 1986), pp. 73-80.
This article presents a thoughtful essay on model management and proposes SM as a foundation for designing a generalized model management system.
95. Dolk, D.R., "Model Management and Structured Modeling: The Role of an Information Resource Dictionary System," Comm. ACM, 31:6 (June, 1988), pp. 704-718. Presented at the Workshop on Structured Modeling, UCLA, 8/86, with a subsequent progress report at the SM session at the TIMS/ORSA Meeting in New Orleans, 5/87.
This article proposes a design for an active information resource dictionary system as a vehicle for incorporating a SM approach to model management into an organization's information resource management environment. The design uses the relational data model, is compatible with Federal standards, and seems ripe for prototype implementation.
An additional point of interest of this article is that it provides a design approach for simple single-user SM systems based on a relational database management system as the host implementation environment (cf. Lenard's paper "Representing Models as Data").
96. Dolk, D.R., "A Model Management System for Combat Simulation," Report NPS-54-86-014, Naval Postgraduate School, 71 pages, 11/86.
This is the final report of a feasibility study done for the U.S. Army TRADOC Systems Analysis Activity. The papers by J.W. Stott and by J.E. Kottemann, both listed later in this Bibliography, were written under subcontract and are overviewed in this paper.
Author's abstract: "The design and implementation of a model management system to support combat modeling is discussed. Structured Modeling is introduced as a formalism for representing mathematical models. A relational information resource dictionary system is developed which can accommodate structured models. An implementation is described. Structured Modeling is then compared to Jackson System Development (JSD) as a methodology for facilitating discrete event simulation. JSD is currently better at representing the dynamic aspects of simulation whereas Structured Modeling excels in representing the static aspects. A structured model of an existing combat model is presented. Finally, recommendations are made to strengthen Structured Modeling as a tool for discrete event simulation."
97. Dolk, D., "Model Management Systems for Operations Research: A Prospectus," in G. Mitra (ed.), Mathematical Models for Decision Support, Springer-Verlag, Berlin, 1988, pp. 347-373. Presented at a NATO Advanced Study Institute in Val d'Isere, France, 1987.
This paper views models as an organizational resource to be shared among model builders, administrators, and end-user clients; and views model management systems as a means for supporting the above over the entire modeling life cycle. It identifies three main functional requirements: (1) model description, for which SM is recommended as a foundation; (2) model manipulation, for which object-oriented programming is recommended; and (3) model control, for which relational DBMS is recommended.
98. Dolk, D., "Structured Modeling and Discrete Event Simulation," Dept. of Administrative Sciences, Naval Postgraduate School, 26 pages, 5/90.
This paper works toward a synthesis of SM and discrete event simulation that would enable integrated modeling systems to combine analytical models with simulation models. It gives particular attention to mathematical programming models, and reworks the central simulation model of Derrick's thesis (see Part IIA of this Bibliography).
98A. Dolk, D., "Integrated Model Management in the Data Warehouse Era," to appear in European Journal of Operational Research, 1999.
Once or twice each decade, it seems that very important commercial developments take place that the model management community can ignore only at its great peril. This paper was sparked by two such developments of the 90s, namely the data warehousing boom and the advent of the Web. It sketches how future model management systems of ambitious capability can be designed so as to leverage these two developments. Key roles are suggested for decision metrics, models as predictors of decision metrics, SM and/or UML for model representation, and a component-based architecture.
99. Dolk, D. and M. Ackroyd, "Enterprise Modeling and Object Technology," Proceedings of the Third ISDSS Conference, The International Society for Decision Support Systems, Volume 2, Hong Kong, 6/95. Available as a working paper, Naval Postgraduate School, 16 pages, 3/95. Another version was presented at the "New Directions in Structured Modeling" session at the INFORMS International Singapore, June 25-28, 1995.
This paper examines five major approaches to enterprise modeling (EM), namely those based on information systems architecture, DSS, discrete event simulation, business process reengineering, and object orientation, and finds them all too narrow. A model management approach is needed. For this the authors recommend embedding SM into Rumbaugh's Object Modeling Technique (OMT) as an enabling implementation technology. Reasons: SM has the modeling generality needed by EM, and permits decision and information models to be integrated. A mapping of SM into OMT is presented.
100. Dolk, D. and J. Kottemann, "Model Integration and a Theory of Models," Decision Support Systems, 9:1 (January 1993), pp. 51-63.
This article discusses model integration from 4 points of view: organizational, definitional, procedural, and implementational. The discussion of the second viewpoint is based largely on SM and ideas relating to schema integration. The discussion of the third viewpoint uses SML for an extended example, and gives a tantalizing glimpse of a model manipulation language based on SML. The discussion of the fourth viewpoint makes extensive use of the O-O paradigm, and briefly describes "Communicating Structured Models" based on communicating sequential processes. The article closes with a crucial research question, namely how to put model interconnection languages on a rigorous basis.
101. Dresbach, S., "Modeling by Construction: A New Methodology for Constructing Models for Decision Support," Proceedings of the Twenty-Ninth Annual Hawaii International Conference on System Sciences, Vol. II, IEEE Computer Society Press, Los Alamitos, CA, pp. 178-197, 1/96. On-line as a PostScript file here.
This paper positions itself as an approach complementary to SM for the early phases of model construction. Using the concept of "independence", which is detailed in a companion paper, the author gives a hierarchy of abstraction levels of model specification that is much richer than SM's model-instance/model-schema distinction.
101A. Dresbach, S., "Modeling Methodologies for Modeling Environments," to appear in P. Kall, I. Lasiecka, and M. Polis (eds.), System Modelling and Optimization, Addison Wesley Longman, 1999. Available as a working paper, Ministry for Science and Research of Northrhine-Westfalia, D-40190 Düsseldorf, Germany, 10 pages, 7/97; on-line here.
This paper gives an extended exposition and critique of SM, together with a briefer exposition of the author's own "modeling by construction" framework. The author suggests that the two frameworks be integrated, with SM serving as the theoretical basis.
102. Eck, R., A. Philippakis and R. Ramirez, "Solver Representation for Model Management Systems," Proceedings of the Twenty-Third Annual Hawaii International Conference on System Sciences, Vol. III, IEEE Computer Society Press, Los Alamitos, CA, pp. 474-483, 1/90. A follow-on paper by the same authors, "The Encoding and Use of Knowledge About Solvers in Model Management Systems," was accepted by Decision Support Systems and then withdrawn.
This paper focuses on the dynamics of model-solver interaction in a model management system under the assumption that SML is used as the model definition language. Interactions of interest include specifying values for variable attributes, evaluation, finding solvers that "match" a given model, optimization, queries, and inferencing. The general approach is to think of solvers (in a broad sense) as objects having 3 components: (a) Preconditions (assertions about model elements and their "states" before solver invocation), (b) PostConditions (assertions about model elements and their "states" after solver invocation), and (c) Prototype (specifications for the model-solver interface). This is an overview paper; many details remain to be given.
103. Gagliardi, M. and C. Spera, "Toward a Formal Theory of Model Integration," in I. Maros and G. Mitra (eds.), Annals of Operations Research, 58 (1995), pp. 405-440.
This paper formalizes some of the mechanics of model integration ("deep" in the sense of item 16 of this Bibliography), working entirely within the formal structured modeling framework. The examples are drawn from item 8 of this Bibliography using BLOOMS (an object-oriented language for SM developed by the authors, listed in a nearby item) rather than SML.
104. Gagliardi, M. and C. Spera, "Some New Results in Model Integration," Proceedings of the Twenty-Eighth Annual Hawaii International Conference on System Sciences, Vol. III, IEEE Computer Society Press, Los Alamitos, CA, pp. 398-407, 1/95.
This article extends the results given by the same authors in "Toward a Formal Theory of Model Integration" (the previous item in this Bibliography).
105. Gagliardi, M. and C. Spera, "BLOOMS: A Basic Language Object Oriented for Modeling Systems," Working Paper 10, Dept. of Quantitative Methods, University of Siena (Italy), 34 pages, 9/95. Earlier version presented at the AIEM3 Conference, Portland, 25-27 August 1993. Further details are given in "The Syntax of BLOOMS," a pair of technical reports by the same authors, most recently revised 2/95.
This paper spells out in detail one possible view of the relationship between SM and the object-oriented approach. It then details the consistency of the authors' new object-oriented model definition language, BLOOMS, with each of the core concepts of SM. BLOOMS, which was influenced by Eiffel, is used by a model management system being developed by the authors. The implementation is in C and C++, with the GUI in Visual Basic.
106. Gagliardi, M. and C. Spera, "BLOOMS: a Prototype Modeling Language with Object Oriented Features," Decision Support Systems, 19:1 (January 1997), pp. 1-21.
This paper explains some of the motivation and strategy behind the design and implementation of the authors' object-oriented language for SM-style model definition.
107. Gazmuri, P. and I. Arrate, "Modeling and Visualization for a Production Planning Decision Support System," International Transactions in Operational Research, 2:3 (1995), pp. 249-258. Presented at IFORS-SPC3 (Digital Technologies/Multimedia: OR/MS in Strategy, Operations and Decision Support), Santa Monica, 1/95. Also available as a working paper, Pontificia Universidad Católica de Chile, Dept. of Industrial and Systems Engineering, 16 pages, 3/29/95.
This paper falls under the project described in the neighboring item by Gazmuri, Maturana, Vicuña and Contesse. It reviews the general architecture of the implementation, which uses PowerBuilder for the graphic user interface, WATCOM SQL for the relational database, CPLEX for mixed integer optimization, and some of FW/SM's code for handling SML. This implementation was released in January 1995 to one of the sponsoring companies -- namely, the largest appliance manufacturer in Chile -- as a practical production planning system. The problems being solved have about 2000 continuous variables, 600 binary variables, and 2800 general constraints.
There is a 62-page User Manual (in Spanish) for this application.
108. Gazmuri, P., S. Maturana, F. Vicuña and L. Contesse, "Design of a Generator of Computer Systems for the Optimization of Productive Processes," proposal by Pontificia Universidad Católica de Chile to FONDEF (Fund for the Advancement of Scientific and Technological Development), 57 pages plus appendices (in Spanish), 4/93. Project Web page: http://ing.puc.cl/~gescopp. See also P. Gazmuri, S. Maturana and L. Contesse, "Una Proposición de Metodología y Herramientas para el Desarrollo de Sistemas de Apoyo a la Gestión Basados en Modelos Analíticos," Apuntes de Ingeniería, 52 (April-June 1994), pp. 71-93.
This rather large project, funded from 1994-1997, used SML as the basis for a "software factory" to generate custom prototypes for selected optimization problems for production and distribution taken from two participating Chilean manufacturing firms. Approaches explored included the one in "Application-Specific Delivery Systems for Optimizer Engines" (item 49 of this Bibliography). The first year of the project focused on developing prototypes by hand. Attempts to automate the whole process as much as possible were the focus of the second year. The third year was devoted to testing and evaluating the results. See the items by Gazmuri and Arrate, by Maturana and Eterovic, by Maturana, Ferrer and Barañao, and by Maturana, Gazmuri and Villena.
There was also a similar but smaller project led by S. Maturana, F. Vicuña and J. Gana, "Methodologies and Tools for Developing DSS Based on Analytic Models." Funded by FONDECYT (National Fund for Scientific and Technological Development of Chile) for two years starting 3/94, it used pieces of FW/SM among other tools for generating user interfaces and for interacting with a relational DBMS.
109. Greenberg, H.J. and F.H. Murphy, "Views of Mathematical Programming Models and Their Instances," Decision Support Systems, 13:1 (Jan 1995), pp. 3-34.
Almost any representation of a model, whether partial or complete, is a "view" as the term is used in this paper. The point is that a judiciously selected view can be more insightful for the purpose or task at hand than the main representation of the full model. This paper discusses many possible views for mathematical programming models, including algebraic, block schematic, graphic, and textual. A detailed discussion of SM is included from this perspective.
It is indeed important for a modeling environment to accommodate many possible views, but it may be that the main representation of the full model ought to contain all essential information needed to construct any view. Then support for multiple views reduces, in the case of a system like FW/SM, to the chore of adding new options to the REFGEN process (and possibly splitting REFGEN into independent processes -- such as NETDRAW -- each of which generates a desired view).
110. Hamacher, S., P. Dejax and L. Lustosa, "A Diagram Representation for Conceptual Models of Operations Research Problems," to appear in International Transactions in Operational Research. Available as Cahiers d'Etudes et de Recherche No. 93-04A, Laboratoire Productique Logistique, Ecole Centrale Paris, France, 16 pages, 5/93. Also available as Technical Memorandum 07/93, Centro Tecnico Cientifico, Departamento de Engenharia Industrial, Pontificia Universidade Católica do Rio de Janeiro, Brazil, 7/93. Presented at the IFORS 93 Conference in Lisbon, 7/93.
The aim of this paper is to move toward a more usable graphical user interface for computer-based modeling environments. It proposes a variant of the standard genus graph that adds semantic data modeling features from the Entity Relationship Model extended to include specialization, generalization, and aggregation. This is done in a SML-compatible way, so that a standard SML schema can be produced from the proposed diagram.
A research prototype called IGOR (Integrated Graphics for Operations Research) uses a graphical object-oriented programming language (GraphTalk) under Windows 3.1. IGOR is quite rich in features useful for model construction and query, and makes essential use of Microsoft Word and Microsoft Access.
The authors have presented several progress reports on IGOR: at the IFORS 93 Conference in Lisbon, 7/93; at the Spring TIMS/ORSA Meeting in Boston, 4/94; and at IFORS-SPC3 (Digital Technologies/Multimedia: OR/MS in Strategy, Operations and Decision Support), Santa Monica, 1/95. They also have a paper "IGOR: Un Systeme de Modelisation Integre" in Proceedings of the International Conference in Industrial Engineering, Wroclaw, Poland, 10/94. There is also a paper by S. Hamacher, L. Lustosa and P. Dejax, “IGOR: Um sistema integrado de modelagem para programação matemática,” Working Paper 06/97 (1997), Departamento de Engenharia Industrial, PUC-Rio, submitted for publication in the Brazilian OR Society's journal Pesquisa Operacional.
See the first author's Ph.D. dissertation earlier in this Bibliography. There is also a 38-slide PowerPoint slide show "IGOR: Integrated Graphical Modeling System for Operational Research," 11/94, available in English or French.
111. Heavey, C., "Structured Modeling - A Basis for a MMS for Manufacturing Systems Design," technical report, Dept. of Manufacturing and Operations Engineering, University of Limerick, Ireland, 52 pages, 7/93.
This paper proposes two extensions of SML to improve its applicability to manufacturing systems design: one facilitating modeling discrete event dynamic systems (a sixth entity type called "decision point entity" for describing dynamic behavior rules, together with a new type of attribute value designed for table-lookup), and another extension allowing genera of two model schemas to be linked for purposes of data transfer. A lengthy appendix contains several queueing network models that use the two extensions.
112. Heavey, C. and J. Browne, "A Model Management Systems Approach To Manufacturing Systems Design," International J. of Flexible Manufacturing Systems, 8:2 (April 1996). Also available as a working paper, Dept. of Manufacturing and Operations Engineering, University of Limerick, Ireland, 35 pages, 7/95.
This paper stresses the applicability of model management systems ideas to manufacturing systems design. Building on the earlier paper by Heavey (in this Bibliography) that extends SML, it illustrates these ideas via a prototype SM implementation using X Window, C++, embedded SQL for storing elemental detail data in a RDBMS, and OSF/Motif. It offers three distinct user interfaces called "workbenches," one for model developers, one for model users, and one for manufacturing system designers. A general purpose solver interface, not unlike that of GENNET in FW/SM, is included to facilitate attaching additional solvers. The prototype was evaluated using queueing network models for manufacturing systems design.
113. Holocher, M., R. Michalski, D. Solte and F. Vicuña, "MIDA: An Open Systems Architecture for the Model-Oriented Integration of Data and Algorithms," Decision Support Systems, 20:2 (June 1997), pp. 135-147. An earlier paper by R. Mueller, F. Ricci, D. Solte and F. Vicuña, "AMBAS-SOOM: A Concept for an Open Systems Architecture for the Model-Based Integration of Data and Algorithms," was presented at the 3rd International Workshop on Data, Expert Knowledge and Decisions (Reisensburg Castle, G uenzburg, Germany, 23-27 September 1991).
This paper sketches the concepts behind a new architecture for decision support systems for distributed heterogeneous computing environments. Such environments inevitably will become increasingly important in the future. The prototype system, which is implemented in C++ under Unix, is called AMBAS-SOOM (Adaptive Method BAse Shell-Structured and Object Oriented Modeling). It combines structured modeling ideas and the SML language with Solte's "operational programming" paradigm for software engineering in object-oriented style.
114. Hua, H. and S.O. Kimbrough, "A Case Study in Model Management: The US Coast Guard, GAO, TEFA-2, and Structured Modeling with Embedded Languages," Proceedings of the Thirtieth Annual Hawaii International Conference on System Sciences, IEEE Computer Society Press, Los Alamitos, CA, 1/97. On-line here as a pdf file and here as a PostScript file). Related earlier work was presented at the "New Directions in Structured Modeling" session at the INFORMS International Singapore, June 25-28, 1995. On-line here.
This paper begins with a summary and thoughtful analysis of modeling criticisms raised in a General Accounting Office audit of an unsuccessful U.S. Coast Guard proposal to acquire a new fleet of patrol boats. The authors respond to these criticisms by adding structured modeling features to the TEFA model management system built previously for the Coast Guard. This is facilitated by exploiting the embedded languages technique pioneered by Kimbrough and his former student H. Bhargava. This technique enables rich information about models to be incorporated and to be easily accessible for machine processing. The internal model representation is not based on SM, but is rich enough to produce SML representations at least for models falling under SML levels 1 and 2. TEFA-2 generates models as documents (e.g., Mathematica notebook files), with the notable advantage that organizations can use existing document processing and document management technologies for dealing with models. Testing has been carried out with several Coast Guard models. TEFA-2 is Prolog-based, runs on both Unix (XWindow) and Macintosh platforms, and relies on Mathematica for solver functionality (although drivers for other solvers should not be difficult to write). TEFA-2 can produce reports as hypertext pages suitable for access by WWW browsers, an innovative idea that helps make models and their results more easily accessible to non-modeling users. Click here for Web access.
115. Huh, Soon-Young, "Version Management Concept in a Model Management System," John E. Anderson Graduate School of Management, UCLA, 32 pages, 10/89.
Version management ideas from CAD and CASE are adapted for use in a modeling environment based on SM. There is a fairly strong object-oriented flavor.
116. Huh, Soon-Young, "Modelbase Construction with Object-Oriented Constructs," Decision Sciences, 24:2 (March-April 1993), pp. 409-434.
This article focuses on the modelbase component of a DSS, with particular attention to logical independence among models, data, and solvers. SML is adopted as the model representation language. Three abstraction layers are employed: model type, model structure, and model instance. The OODB concepts of inheritance, instantiation, encapsulation, and polymorphism play a major role. A prototype implementation has been developed using C++ and the OODBMS ObjectStore on a Sun-4. Version management features and a graphic user interface are likely future enhancements.
117. Huh, Soon-Young and Q.B. Chung, "A Model Management Framework for Heterogeneous Algebraic Models: Object-oriented Database Management Systems Approach," Omega, 23:3 (1995), pp. 235-256.
This article continues the work begun in Huh's dissertation (an item in Part IIA of this Bibliography) and his Decision Sciences article (see the preceding item). The three main research objectives are: (a) to accommodate simultaneously different algebraic modeling languages (AMPL, GAMS, and SML are given special attention), (b) to exploit the capabilities of an object-oriented database management system (ObjectStore is chosen), and (c) to devise an integrated language for manipulating data, models, and solvers based on an object-oriented structured query language. An X Window interface is being added to the C++ prototype.
117A. Huh, Soon-Young, Hyung-Min Kim and Q.B. Chung, " Framework for Change Notification and View Synchronization In Distributed Model Management Systems," Working Paper, KAIST, Taejon, Korea, 25 pages, 1998.
This paper proposes a change notification framework where multiple group users can collaborate by sharing underlying models with different views but maintaining synchronized views over evolving models. The framework is to support view synchronization in distributed model management systems that (1) manage the dependency relationships between the shared models and the dependent user views, and (2) coordinate the changes by way of controlling and communicating the propagation activities between the two. The framework identifies the change requirements based on the locus of changes in the shared models and the effects of such changes in the dependents, and further defines the change management constructs and processes to particularize the framework into a view synchronization mechanism for a distributed modeling environment. In the framework, Structured Modeling was adopted for the representation of mathematical models, the generic model concept for constructing and dealing with mathematical models, and the object-oriented database (ODBMS) paradigm for combining the constructs in a single formalism. A prototype system has been developed at KAIST using a commercial ODBMS called ObjectStore and C++.