Decision-Making Support Systems

Transcription

1 978 Category: Software & Systems Design Decision-Making Support Systems Guisseppi Forgionne University of Maryland, Baltimore County, USA Manuel Mora Autonomous University of Aguascalientes, Mexico Jatinder N. D. Gupta University of Alabama-Huntsville, USA Ovsei Gelman National Autonomous University of Mexico, Mexico IntroductIon Decision-making support systems (DMSS) are computerbased information systems designed to support some or all phases of the decision-making process (Forgionne, Mora, Cervantes, & Kohli, 2000). There are decision support systems (DSS), executive information systems (EIS), and expert systems/knowledge-based systems (ES/KBS). Individual EIS, DSS, and ES/KBS, or pair-integrated combinations of these systems, have yielded substantial benefits in practice. DMSS evolution has presented unique challenges and opportunities for information system professionals. To gain further insights about the DMSS field, the original version of this article presented expert views regarding achievements, challenges, and opportunities, and examined the implications for research and practice (Forgionne, Mora, Gupta, & Gelman, 2005). This article updates the original version by offering recent research findings on the emerging area of intelligent decision-making support systems (IDMSS). The title has been changed to reflect the new content. 2002; Turban & Aronson, 1998) such as management support systems (MSS), decision technology systems (DTS), integrated DMSS, data warehouse (DW)-based and data mining (DM)-based DMSS (DW&DM-DMSS), intelligent DMSS (i-dmss), and Web-based DMSS or knowledgemanagement DMSS. The architectures have been applied to various public and private problems and opportunities, including the planning of large-scale housing demand (Forgionne, 1997), strategic planning (Savolainen & Shuhua, 1995), urban transportation policy formulation (Rinaldi & Bain, 2002), health care management (Friedman & Pliskin, 2002), pharmaceutical decision making (Gibson, 2002), banking management (Hope & Wild, 2002), entertainment industry management (Watson & Volovino, 2002), and military situations (Findler, 2002). Applications draw on advanced information technologies (IT), such as intelligent agents (Chi & Turban, 1995), knowledge-based (Grove, 2000) and knowledge-management procedures (Alavi, 1997), synthetic characters (Pistolesi, 2002), and spatial decision support systems (Silva, Eglese, & Pidd, 2002), among others. Background Decision-making support systems utilize creative, behavioral, and analytic foundations that draw on various disciplines (Sage, 1981). These foundations give rise to various architectures that deliver support to individual and group DMSS users. The architectures, which are summarized in Table 1, include (a) classic systems (Alter, 1996) such as decision support systems (DSS), expert and knowledge-based systems (ES/KBS), executive information systems (EIS), group support systems (GSS), and spatial decision support systems (SDSS) and (b) new systems (Forgionne, 1991; Forgionne, Mora, Cervantes, & Gelman, 2002a; Gray & Watson, 1996; Mora, Forgionne, Gupta, Cervantes, & Gelman, 2003; Power, dmss achievements Once created, DMSS must be evaluated and managed. Economic-theory-based methodologies, quantitative and qualitative process and outcome measures, and the dashboard approach have been used to measure DMSS effectiveness. These approaches suggest various organizational structures and practices for managing the design, development, and implementation effort. Most suggestions involve much more user involvement and a larger role for nontraditional specialists during the technical design, development, and implementation tasks. To gain further insights about DMSS achievements, challenges, and opportunities posed by the development, the Copyright 2009, IGI Global, distributing in print or electronic forms without written permission of IGI Global is prohibited.

2 Table 1. Decision-making support systems architectures D Classic DMSS Architectures Description Main Decision-Making Phase Supported INTELLIGENCE DESIGN CHOICE IMPLEMENTATION LEARNING DMSS SUPPORT CHARACTERISTICS DSS ES & KBS EIS GSS SDSS A DSS is an interactive computer-based system composed of a user-dialog system, a model processor and a data management system, which helps decision makers utilize data and quantitative models to solve semi-structured problems. An ES/KBS is a computer-based system composed of a user-dialog system, an inference engine, one or several intelligent modules, a knowledge base, and a work memory, which emulates the problem-solving capabilities of a human expert in a specific domain of knowledge. An EIS is a computer based system composed of a user-dialog system, a graph system, a multidimensional database query system and an external communication system, which enables decision makers to access a common core of data covering key internal and external business variables by a variety of dimensions (such as time and business unit). A GSS an integrated computer based system composed of a communication sub-system and model-driven DMSS (DSS), to support problem formulation and potential solution of unstructured decision problems in a group meeting. A SDSS a computer based system composed of a user-dialog sub-system, a geographic/ spatial database sub-system, a decision model sub-systems and a set of analytical tools, which enables decision makers to treat with situations based strongly on spatial data. A (A) What-if, goal seeking, & sensitivity analysis. A B (A&B) Symbolic pattern-based recognition; fuzzy data; how and why explanation facilities. A B (A&B) Key performance indicators (KPI s) in graphs and text tables; data exploring and searching through drill-down, roll-up, slice and dice and pivoting operations; networking communications to internal and external bulletin boards. A B (A) Idea generation through brainstorming facilities; pooling and display of ideas; generation of alternatives and criteria. (B) Preference models; voting schemes; conflict negotiation support. A B (A) Spatial data searching support; visualization tools for maps, satellite images, and digital terrains. (B) What-if analysis of scenarios, goal-seeking analysis, sensitivity analysis of decision variables upon spatial data. continued on following page 979

3 Table 1. continued Description Main Decision-Making Phase Supported DMSS SUPPORT CHARACTERISTICS Modern DMSS Architectures INTELLIGENCE DESIGN CHOICE IMPLEMENTATION LEARNING MSS, DTS or I-DMSS DW & DM DMSS Web- DMSS & KM- DMSS i-dmss These systems are the result of the triple-based integration (i.e., DSS, EIS, and ES/KBS) and have the aim to offer a full support to decision maker in all phases of the DMP. DW&DM-DMSS are computerbased system composed of a userdialog sub-system, a multidimensional database subsystem, and an on-line analytical processing (OLAP) component enhanced with knowledge discovery algorithms to identify associations, clusters, and classifications rules intrinsic into the data warehouse. Web-DMSS & KM-DMSS are computer-based system composed of an user-dialog sub-system, a text &multimedia document storage subsystem and publishing/retrieval subsystem to preserve and distribute knowledge in the organization using intranets. Are computer based system composed of an user-dialog sub-system, a multidimensional database and knowledge base subsystem and a quantitative & qualitative processing sub-system enhanced all of them with AI-based techniques, designed to support all phases of the DMP. A B C D (A&D) Visual data exploring through graphs; color codes and tables; data exploration with drill-drown, roll-up, slice, and dice, pivoting operations. (B) (C) Intelligent advice through AI-based capabilities to support the models selection task. Numerical modeling through available numerical-based models; what-if, goal seeking and sensitivity analysis. A (A) OLAP capabilities of aggregation, slice and dice; drill-down; pivoting; trend analysis; multidimensional query; graphics and tabular data support. Knowledge discovery patterns using statistical based, tree-decision or neural networks. A B (A&B) Document publishing and retrieval facilities A B C D E (A&D) Visual data exploring through graphs; color codes and tables; data exploration with drill-drown, roll-up, slice, and dice, pivoting operations. (B) (C) (E) Intelligent advice through AI-based capabilities to support the models selection task. Numerical and qualitative modeling through numerical-based or symbolic models; what-if, goal seeking, and sensitivity analysis. Symbolic reasoning through knowledgebased models for explanations about how and why the solution was reached. 980

4 Table 2. DMSS achievements, challenges, and opportunities D DMSS Issue Key Achievements Research Issues and Practical Problems Core DMSS Architectural Concepts and Opportunities Expert Collective Opinion The evolution of DMSS software and hardware; the implementation of DMSS in a variety of organizations; the creation of DMSS tailored design and development strategies Providing quality data for decision support; managing and creating large decision support databases; model management and model reuse; building knowledge driven DMSS; improving communication technologies; developing a uniform and comprehensive DMSS scheme; developing an effective toolkit; developing and evaluating a synergistic integrated DMSS; collecting insights about the neurobiology of decision support for managers less structured work; the application of agent and object-oriented methodologies; developing DMSS though well-established methodologies Web technology; accessibility; security; effective data, idea, and knowledge management, possibly through the use of smart agents; effective model management; effective dialog management; EISlike features; incorporation of basic and common DMSS functionalities; mobile computing; usercentric design. original study compiled opinions from recognized leaders in the field (Forgionne, Gupta, & Mora, 2002b). The expert verbatim views are summarized in Table 2. expert opinions The expert opinion indicates that DMSS have been recognized as unique information systems. Collectively, these experts focus on the deployment of new and advanced information technology (IT) to improve DMSS design, development, and implementation. In their collective opinion, the next generation of DMSS will involve: (a) the use of portals, (b) the incorporation of previously unused forms of artificial intelligence through agents, (c) better integration of data warehousing and data mining tools within DMSS architectures, (d) creation of knowledge and model warehouses, (e) the integration of creativity within DMSS architectures, (f) the use of integrated DMSS as a virtual team of experts, (g) exploitation of the World Wide Web, (h) the exploitation of mobile IT, and (i) the incorporation of advanced IT to improve the user interface through video, audio, complex graphics, and other approaches. Future opportunities, trends and challenges discerned by the experts include: (a) availability of DMSS packages for specific organizational functions, such as customer relationship management, (b) system functional and technical integration, consolidation, and innovation, (c) software tool cost education, (d) the creation of a technology role for the decision maker through the DMSS, (e) the integration of the decision maker into the design and development process, (f) developing effective design and development tools for user-controlled development, (g) accommodating the structural changes in the organization and job duties created by DMSS use, (h) developing new and improved measures of DMSS effectiveness, (i) incorporating the cognitive and group dimensions of decision making, (j) utilization of smart agents, (k) distribution of DMSS expertise through collaborative technologies, (l) incorporating rich data, information and knowledge representation modes into DMSS, and (m) focusing user attention on decisions rather than technical issues. Common themes suggested by this disparate expert opinion are (a) the DMSS should focus decision makers on the decision process rather than technical issues, and (b) DMSS development may require specialized and new IT professionals, and (c) there is need for a systematic and well-managed implementation approach. Intelligent dmss Since most experts value artificial intelligence in decision making support, a historical review of the literature, covering the period , was conducted to examine the state of the intelligent DMSS (I-DMSS) concept (Mora et al., 2006). This history indicated that neural networks and fuzzy logic have become more popular than Bayesian/belief nets, and intelligent agents, genetic algorithms, and data mining have emerged as tools of interest. In terms of the decision making process, the intelligence and choice phases have been the most supported phases. Over time, intelligence support has increased, while choice support has decreased. Within the intelligence phase, the problem recognition step has grown in popularity. Among dialog user interface capabilities, text/passive graphics has remained the most used tool. Model management has been most often supported by knowledge-based methodologies and quantitative models. Knowledge-based models have been declining in importance, while quantitative models have been gaining popularity. Symbolic structured mechanisms, based on rule-based systems and fuzzy logic, and quantitative structured approaches, based on neural networks and data mining, have become the most popular data management tools. 981

5 future trends The historical analysis supports some of the expert opinion. Specifically, the reported record indicates that effort is underway to (a) increase DMSS processing capabilities through intelligent agents, fuzzy systems, and neural networks and (b) improve user-interface capabilities through multimedia and virtual environments. In short, the experts and literature on AI and DMSS implicitly recognize the relevance of improving the DMSS user interface, information and knowledge representations schemes and intelligent processing capabilities through the deployment of advanced IT. conclusion In some ways, the DMSS field has not progressed very much from its early days. There is still significant disagreement about definitions, methodologies, and focus, with expert opinion varying on the breadth and depth of the definitions. Some favor analytical methodologies, while others promote qualitative approaches. Some experts focus on the technology, while others concentrate on managerial and organizational issues. There does not seem to be a unified theory of decision-making, decision support for the process, or DMSS evaluation. Moreover, achieving successful implementation of large-scale DMSS is still a complex and open research problem (Mora et al., 2002). In spite of the diversity, opinions are consistent regarding some key DMSS elements. Most experts recognize the need for problem pertinent data, the role of the Internet in providing some of the necessary data, the need for system integration within DMSS architectures and between DMSS and other information systems, and the importance of artificial intelligence within DMSS processing. The historical record also supports the emerging importance of intelligent decision-making support and identifies quantitative-based methodologies as the growing form of intelligence. The DMSS concept also continues to be successfully applied across a variety of public and private organizations and entities. These applications continue to involve the user more directly in the design, development, and implementation process. The trends will create DMSS that are technologically more integrated, offer broader and deeper support for decisionmaking, and provide a much wider array of applications. In the process, new roles for artificial intelligence will emerge within DMSS architectures, new forms of decision technology and methodology will emerge, and new roles will be found for existing technologies and methodologies. As the evolution continues, many tasks that had been assigned to human experts can be delegated to virtual expertise within the DMSS. With such consultation readily available through the system, the decision maker can devote more effort to the creative aspects of management. Support for these tasks can also be found within DMSS. In the process, the decision maker can become an artist, scientist, and technologist of decision-making. The DMSS-delivered virtual expertise can reduce the need for large support staffs and corresponding organizational structures. The organization can become flatter and more project-oriented. In this setting, the decision maker can participate more directly in DMSS design, development, implementation, and management. Such changes will not occur without displacements of old technologies and job activities, radical changes in physical organizations, and considerable costs. As the reported applications indicate, however, the resulting benefits are likely to far outweigh the costs. references Alavi, M. (1997). KPMG Peat Marwick U.S.: One giant brain. In Creating a system to manage knowledge (pp ). Harvard Business School Publishing (Case ). Alter, S. (1996). Information systems: A management perspective. Menlo Park, CA: Benjamin/Cummings. Chi, R., & Turban, E. (1995). Distributed intelligent executive information systems. Decision Support Systems, 14(2), Findler, N. (2002). Innovative features in a distributed decision support system based on intelligent agent technology. In M. Mora, G. Forgionne, & J. Gupta (Eds), Decision-making ). Hershey, PA: Idea Group Publishing. Forgionne, G. (1997). HADTS: A decision technology system to support army housing management. European Journal of Operational Research, 97(2), Forgionne, G. (1991). Decision technology systems: A vehicle to consolidate decision-making support. Information Processing and Management, 27(6), Forgionne, G., Mora, M., Cervantes, F., & Gelman, O. (2002a, July 3-8). I-DMSS: A conceptual architecture for next generation of DMSS in the Internet age. In F. Adam, P, Brezillon, P. Humpreys, & J. Pomerol (Eds.), Proceedings of the International Conference on Decision Making and Decision Support in the Internet Age (DSIAge02) (pp ). Cork, Ireland. Forgionne, G., Mora, M., Cervantes, F., & Kohli, R. (2000, August 10-13). Development of integrated decision-making support systems: A practical approach. In M. Chung (Ed.), Proceedings of the AMCIS 2000 Conference (pp ). Long Beach, CA, USA. 982

6 Forgionne, G., Gupta, J., & Mora, M. (2002b). Decision making support systems: Achievements, challenges, and opportunities. In M. Mora, G. Forgionne, & J. Gupta (Eds.), Decision-making support systems: Achievements, challenges, and trends (pp ). Hershey, PA: Idea Group Publishing. Forgionne, G., Mora, M., Gupta, J. N. D., & Gelman, O. (2005). Decision-making support systems. In M. Khosrow- Pour (Ed.), Encyclopedia of information science and technology (pp ). Hershey, PA: Idea Group Publishing. Friedman, N., & Pliskin, N. (2002). Demonstrating valueadded utilization of existing databases for organizational decision-support. Information Resources Management Journal, 15(4), Gibson, R. (2002). Knowledge management support for decision making in the pharmaceutical industry. In M. Mora, G. Forgionne, & J. Gupta (Eds), Decision-making support systems: Achievements, challenges, and trends (pp ). Hershey, PA: Idea Group Publishing. Glass, R., Ramesh, V., & Vessey, I. (2004). An analysis of research in computing discipline. Communications of the ACM, 47(6), Gray, P., & Watson, H. (1996, August 16-18). The new DSS: data warehouses, OLAP, MDD, and KDD. Proceedings of the AMCIS Conference Phoenix, AZ, USA. Grove, R. (2000). Internet-based expert systems. Expert Systems, 17(3), Hope, B., & Wild, R. (2002). Procedural cuing using expert support system. In M. Mora, G. Forgionne, & J. Gupta (Eds.), Decision-making support systems: Achievements, challenges, and trends (pp ). Hershey, PA: Idea Group Publishing. Mora, M., Cervantes, F., Gelman, O., Forgionne, G., Mejia, M., & Weitzenfeld, A. (2002). DMSS implementation research: A conceptual analysis of the contributions and limitations of the factor-based and stage-based streams. In M. Mora, G. Forgionne, & J. Gupta, (Eds.), Decision-making ). Hershey, PA: Idea Group Publishing. Mora, M., Forgionne, G., Gupta, J., Cervantes, F., & Gelman, O. (2003, Sep. 4-7). A framework to assess intelligent decision-making support systems. In V. Palade, R. Howlett, & L. Jain (Eds.), Proceedings of the 7 th KES2003 Conference, Oxford, UK, LNAI 2774 (pp ). Heiderberg, FRG: Springer-Verlag. Mora, M., Forgionne, G., Gupta, J. N. D., Garrido, L., Cervantes, F., & Gelman, O. (2006). A strategic descriptive review of intelligent decision-making support systems research: The Period. In J. N. D. Gupta, G. A. Forgionne, & M. Mora (Eds.), Intelligent decision-making support systems: Foundations, applications, and challenges, series: Decision engineering (pp ). Springer. Pistolesi, G. (2002). How synthetic characters can help decision-making. In M. Mora, G. Forgionne, & J. Gupta (Eds.), Decision-making support systems: Achievements, challenges, and trends (pp ). Hershey, PA: Idea Group Publishing. Power, D. (2002). Categorizing decision support systems: A multidimensional approach. In M. Mora, G. Forgionne, & J. Gupta (Eds.), Decision-making support systems: Achievements, challenges, and trends (pp ). Hershey, PA: Idea Group Publishing. Rinaldi, F., & Bain, D. (2002). Using decision support systems to help policy makers cope with urban transport problems. In M. Mora, G. Forgionne, & J. Gupta (Eds.), Decision-making ). Hershey, PA: Idea Group Publishing. Sage, A. (1981). Behavioral and organizational considerations in the design of information systems and process for planning and decision support. IEEE Transactions on Systems, Man, and Cybernetics, 11(9), Savolainen, V., & Shuhua, L. (1995). Strategic decision-making and intelligent executive support system. In Proceedings of the 12 th International Conference on Systems Science (pp ), Wroclaw, Poland. Silva, F., Eglese, R., & Pidd, M. (2002). Evacuation planning and spatial decision making: Designing effective spatial decision support systems through integration of technologies. In M. Mora, G. Forgionne, & J. Gupta (Eds.), Decision-making ). Hershey, PA: Idea Group Publishing. Turban, E., & Aronson, J. (1998). Decision support systems and intelligent systems (pp ). Upper Saddle River, NJ: Prentice-Hall. Watson, H., & Volonino, L. (2002). Customer relationship management at Harrah s Entertainment. In M. Mora, G. Forgionne, & J. Gupta (Eds.), Decision-making support systems: Achievements, challenges, and trends (pp ). Hershey, PA: Idea Group Publishing. key terms Data Warehousing-Data Mining (DW-DM) DMSS: Computer-based system composed of an user-dialog subsystem, a multidimensional database subsystem, and an D 983

7 online analytical processing (OLAP) component enhanced with knowledge discovery algorithms to identify associations, clusters, and classifications rules intrinsic in a data warehouse. Decision Making Support System (DMSS): An information system designed to support some, several or all, phases of the decision making process. Decision Support System (DSS): An interactive computer-based system composed of a user-dialog system, a model processor and a data management system, which helps decision makers utilize data and quantitative models to solve semi-structured problems. Executive Information System (EIS): A computer based system composed of a user-dialog system, a graph system, a multidimensional database query system and an external communication system, which enables decision makers to access a common core of data covering key internal and external business variables by a variety of dimensions (such as time and business unit). Expert System/Knowledge Based System (ES/KBS): A computer-based system composed of a user-dialog system, an inference engine, one or several intelligent modules, a knowledge base and a work memory, which emulates the problem-solving capabilities of a human expert in a specific domain of knowledge. Group Support System (GSS): An integrated computer based system composed of a communication sub-system and model-driven DMSS (DSS), to support problem formulation and potential solution of unstructured decision problems in a group meeting. Intelligent Decision Making Support Systems (i- DMSS): Computer based system composed of an user-dialog sub-system, a multidimensional database and knowledge base subsystem, and a quantitative and qualitative processing sub-system enhanced with AI-based techniques, designed to support all phases of the decision making process. Management Support Systems (MSS), Decision Technology Systems (DTS), or Integrated Decision Making Support Systems (I-DMSS): Systems that integrate DSS, EIS and ES/KBS to offer full support to the decision maker in all phases of the decision making process. Spatial Decision Support System (SDSS): A computer-based system composed of a user-dialog sub-system, a geographic/spatial database sub-system, a decision model sub-systems and a set of analytical tools, which enables decision makers to analyze situations involving spatial (geographic) data. Web-DMSS & Knowledge Management (KM)-DMSS: Computer-based system composed of an user-dialog subsystem, a text and multimedia document storage subsystem, and publishing/retrieval subsystem to preserve and distribute knowledge in intranet-supported organizations. 984