Back to search

National Technical University of Athens (NTUA)

Ενοποιημένο περιβάλλον διαχείρισης μοντέλων μαθηματικού προγραμματισμού και ανάπτυξης εφαρμογών υποστήριξης αποφάσεων

Abstract

dc:description

A variety of computer-based modelling environments have been developed in the Operations Research/Management Science (OR/MS) field, to support decision-making when facing real problems. These tools are falling into two main categories; model management systems [6] or integrated modelling environments [35], and Decision Support Systems. To facilitate modelers and analysts to formalise a model for a real problem, model management systems have been developed using high-level symbolic languages for model formulation and running. These languages are often called Algebraic Modelling Languages (AMLs) [40]. GAMS [7], AMPL [16],SML [19], AIMMS [5], MPL [28], LINGO [12] are the most widely used AMLs in modem model management systems. The main advantage of these systems is that they support the model analyst (the modeller) in all stages of the modelling life cycle (table 1-1 in page 3). However, their major disadvantage is that they are not abstract enough to be used by a decision maker, because they require knowledge of the underlying modelling language and the operation of model solver. In order to allow non-expert users to use these tools, an alternative approach is needed. Towards this direction, Huh and Chung [22] proposed a model management framework that provides universal models manipulation commands in order to accommodate models written in different algebraic modelling languages. Geoffrion [20] proposed the concept of structure modelling, which intends to provide a formal mathematical framework, language, and computer environment for conceiving, representing, and manipulating a wide variety of models. Chari and Sen [9] presented a graph-based modelling system for structured modelling that supports model formulation, maintenance and solution using directed acyclic graphs for the representation of the main model entities and their interconnections. Moreover, lijima [23] considered the design and implementation of a system that automatically specifies an optimization model from an abstract form with a given data set specified by the user, and solves it by calling the appropriate solvers. MoGLI [43] is a complete interactive graphic basedmodelling environment for the process of developing, solving and analyzing linear, integer and goal mathematical programmes. Other examples on this approach are PDM [27], and MODFORM [37]. On the other hand the Decision Support Systems (DSS) are standing in a higher abstraction level than the model management systems, and assist the decision maker in the decision-making process. These systems usually integrate a substitute or a simple form of a high level modelling language, models, solvers, database management procedures and a user interface into one package that appears to the user as a single decision support application. Other features provided by these tools are report generation, scenario management, connection to external databases etc. [39]. The DSS are distinguished into two main categories: the data-driven DSS and the model-driven DSS. The data driven DSS emphasize in accessing and manipulating data, and are useful for routine managerial activities such as producing various reports, answering ad-hoc queries, presenting information in different forms, etc. On the other hand, the model-driven DSS emphasise in accessing and managing models and they do not require large and complex databases. However many model-driven DSS incorporate data management procedures and provide OLAP capabilities. The design procedure of model-driven DSS is complex, ill structured and remains essentially ad-hoc. The complexity is result of the need to model not only the problem data and processes, but also themathematical relationships, the integration of data and models, and the decision making style of the decision maker. Several methodologies have been proposed to construct a model-driven DSS that is generic enough to integrate mathematical models of different types for solving various decision problems. Indicative of the significant research that has been conducted on this field is Geoffrion [18] who presented FW/SM, which was a prototype structured modelling environment, and Wright et al. [49] who presented the OR/SM influenced by the work of Geoffrion's and introduced a computerized modelling environment based on structured modelling. Muhanna [32] proposed an object-oriented framework for model management and DSS development, integrating Geoffrion's Structured Modelling approach with Muhanna and Pick's system approach [31]. Raghunathan [36] proposed a methodology based on structured modelling to model and design a DSS. This methodology integrates the research in database design and in the mathematical model formulation within the structured modelling framework.Holocher et al. [21] presented the MIDA architecture suitable for building distributed DSS by integrating high level modelling approaches and implementation techniques for platform independent distributed client/server applications. A Decision Support Framework supporting the PRIMES Energy Model [14], which is based on the GAMS algebraic modelling language and generates a common data structure for models written in GAMS, was presented by Antoniou and Capros [2]. Finally, [17, 34] presented a survey on available commercial model management and model-driven DSS. One of them is AIMMS 3 [42], an all-round development environment for building advanced decision support applications. The development environment includes tools for building a sophisticated graphical end-user interface (GUI), which is highly integrated with the modelling environment. The purpose of this thesis is to describe the design and the features of a model-driven DSS prototype developed at the National Technical University of Athens, using GAMS as the main modelling language, the X-MOD (experience MODelling). The X-MOD combines the advantages of both model management systems and model-based DSS, operating in two different modes in order to support two primary user groups: modelers (who will also be called application developers) and decision makers (who will also be called end-users). In the first mode it operates as a model management system that offers an integrated modelling environment for formulating, developing, solving, and maintaining models written in GAMS. It supports the modeller during the whole modelling life cycle, and it is able to incorporate all kinds of mathematical programming models that are supported by GAMS. The transition to the second operating mode is accomplished through the integrated development environment for building all-round decision support applications, which are ready to run. The ready-torun application incorporates mathematical models, and operates as a specific model-based DSS for a specific decision support problem. It can be used as an alternative user interface for the prototype, or delivered to the decision maker as an executable application, or, finally, published on a web site. Hence in the second operating mode, the prototype runs as a generic model-based DSS that provides the decision makers with an environment focused on decision-making process, designed to meet their individual decision-making styles. The decision environment has been built from inside the prototype, and the decision makers act as end-users who are using the decision support application for a specific problem. From the above is clear that a modeler plays two roles inside the prototype. The first role is the role of the model analyst in which the modeler uses the integrated modelling environment, in order to develop, solve and maintain mathematical models. The second role is the role of the application developer in which the modeler uses the integrated development environment in order to design and implement a ready-to run decision support application for a specific decision problem. So, the prototype operates as a model management system, as a generic model-driven DSS and as an integrated development environment for building ready-to-aim decision support applications. Thus, it stands in the top level of the OR/MS tools hierarchy (Figure 2-2 in page 38), which represents the various OR/MS tools according to the kinds of users that intend to support and to their abstraction of model and solver knowledge. The key motivating factor for designing the prototype is that currently GAMS language lacks support of an integrated system that: a) manages complex and decision-oriented models, b) organizes interrelated multidimensional data, c) employees an advanced graphical user interface builder in order to provide an insight into the solution and problem analysis, and d) operates in a local computer orthrough the Internet/Intranet. It is important to note that the features and components of AIMMS 3 significantly influenced the design and the development of the prototype. However, the X-MOD excels AIMMS 3 in the following points: a) it supports the development of decision-support applications that incorporate more than one models, b) the all-round development environment for decision-support applications can be further programmed using common Active Scripting Languages, c) the task of building a decision-support application is further facilitated by using specialized visual tools, d) the decision-support applications can be either distributed to the decision makers as single executables packages that contain models and data or accessed through an Internet Browser, and e) the open architecture of X-MOD allows the incorporation of third-party developed objects for building a decision support application, and supports the plug-in technology. The structure of the summary is as follows. Section 1 presents the design requirements describing the desired system functionality. In Section 2 we illustrate briefly some of the key components designed to carry out the tasks outlined in section 1 and a case study is examined in Section 3. Finally, Section 4 outlines some suggestions for future research and development. [...]

Degree

thesis:*
Grantor dc:publisher
National Technical University of Athens (NTUA)
Year dc:date
2002

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Πάνος, Ευάγγελος
  • Panos, Evangelos

Subjects

dc:subject × 14

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/13292

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Πάνος, Ευάγγελος; Panos, Evangelos. Ενοποιημένο περιβάλλον διαχείρισης μοντέλων μαθηματικού προγραμματισμού και ανάπτυξης εφαρμογών υποστήριξης αποφάσεων. National Technical University of Athens (NTUA), 2002. http://hdl.handle.net/10442/hedi/13292