Professional Documents
Culture Documents
Decision Support Systems: A Summary, Problems, and Future Trends
Decision Support Systems: A Summary, Problems, and Future Trends
• They agtempt to combine the use of model'g or word decision in DSS implies problem solving.
analytic techniques with traditional data access Indeed, decision and problem solving go hand-in-
and retrieval functions• hand - in making decisions, one is solving prob-
• They specifically focus on features that make lems; conversely, in solving problems, one is mak-
them easy to use by noncomputer people in an ing decision at each step. Furthermore, problem
interactive mode. solving implies the use of knowledge in the pro-
• They emphasize flexibility and adaptability to cess of solving problem. Hence, applications of
accommodate changes in the environment and knowledge, in the form of models, need to be
decision-making approach of the user. incorporated into DSSs. To be effective, it is
essential to know whether a decision to be made
To this list, we add another characteristic of DSS: will be made by an individual or a group based on
single or multiple criteria.
• They support but do not replace upper-level The support aspect of DSS implies the use of
managers in decision making. computer and software technologies to support
managers during the process of decision making.
This characteristic is important as it differentiates Hence it is necessary to understand preferences of
DSS from Expert System. people while they are making decisions. There are
Another approach taken by King (1983) is to four types of support that can be provided to
characterize DSS by a list of components that any decision makers, and are summarized below:
integrated DSS should have:
(1) Passive Support. Provide decision makers with
(a) decision models, DSS that they are comfortable with to allow
(b) interactive computer hardware and software, them to make autonomous decisions.
(c) a data base, (2) Traditional Support: Provide DSSs to decision
(d) a database management system, makers to mesh with the decision and its
(e) graphical and other sophisticated displays, improvements.
(f) a modelling language that is 'user friendly'. (3) Extended Support: Alternatives are actively
suggested by DSSs to decision makers.
As the types of problems to be solved, domains of (4) Normative Support: DSSs dominate the whole
applicable knowledge, and types of targeted users decision process, and managers are there to
are not specified, King's (1983) set of DSSs is provide inputs and specifications.
larger than the set of DSSs normally perceived by
others (Keen and Scott Morton 1978, Huff, 1985; The word system in DSS deviates from the
Hogue, 1985; Meador, Guyote and Rosenfeld, meaning of system used in Checkland (1981). In
1986) in the area of information systems• For the context of DSS, system implies a system of
example, computer-assisted medical diagnosis and man-machine interactions and its design and
town planning are also included in King's set of implementation• An integrated approach, in this
DSSs• case, is important, especially if the system is to
link with existing databases using computer net-
works•
3. Decision, Support and Systems
Commenting on the development of DSS in the 4. Types of Problems and Decision Style
last ten years, Keen (1987) suggested the needs to
adopt a balanced approach to the three aspects of Generally speaking, there are three types of
DSS - namely, Decision, Support and Systems. problems to be solved by management: structured,
This is in response to the recent swing towards send-structured, and unstructured. The boundaries
DSS building. between them, however, are not as clear cut as
Although many management scientists use the they sound• Sometimes the classification is also
term DSS without qualification, it is clearer to call dependent on the existence of methods for solving
it Management Decision Support System. The the problems• For instance, before the advent of
358 M.C Er / Decision Support Systems
linear programming, profit maximization or cost ment control, and strategic planning, which are
minimization was considered as a semi-structured performed by supervisory management, middle
problem; nowadays, with the availability of sim- management, and top management, respectively.
plex method, the problem is generally reduced to a In general, supervisory management deals more
structured problem. with structured problems affected by internal fac-
The levels of management activities, generally, tors, top management more with unstructured
can be classified into operational control, manage- problems affected by external factors, with middle
T
Determine the key I
decisions Define
objectives for
support effort
T
Define "normative"
model(s)
Identify
resources
available
In an earlier attempt, Keen and Scott Morton unsolved problem. Many DSSs are not com-
(1978) proposed a predesign cycle for developing patible with each other, forcing decision
DSS (fig. 5). Indeed, if one knows where the makers to retype data and ~hus creating un-
problems is, one has solved half of the problem necessary duplication of data and efforts.
(Landry, Pascot and Briolat, 1985). For really Problems also arise from the non-compatibil-
unstructured problem, it is generally hard to pin- ity of purchased DSSs with existing databases
point where the problem is. Perhaps the use of and computer networks.
problematique is a better word for describing the (2) Data integrity and secutiry
situation, rather than problem or problematical. Duplication of data poses a data integrity
problem whereby data are at different cycles
of update. Furthermore, the scattering of sen-
6. DSS Applications sitive data at different places poses the secur-
ity problem which is hard to control.
Hitherto, most DSSs are applied to structured (3) Unstructured problems
and semi-structured problems. A list of existing There is a need to extend DSSs to solve un-
DSS applications, non-exhaustive of course, is structured problems which are et,.mmonly
given below: faced by top executives who ma~:e vital deci-
Corporate financial planning sions.
- loan amortization (4) Management of DSSs
- depreciation With small DSSs built and forgotten, and
- lease versus buy large DSSs constantly under modifications
- discounted cash flow and net-present value without documentations, there is a real need
- break-even analysis to manage the process and product of DSS
• Marketing analysis development. Otherwise the whole thing is
- forecasting running out of control, especially when key
- sales analysis developers resign.
- promotion analysis (5) Cost-effectioeness
- consumer sales audits Is a DSS cost effective? Clearly some justifica-
• Real estate investments tions are needed before committing personnel
- financing alternative and monetary resources to the development of
- cash flows a DSS. The traditional cost-benefit analysis
- impact on taxes may not be the best approach for assessing the
- payoff benefit of an innovation• Keen (1981) pro-
• Mineralogical exploration posed the value analysis which stresses value
• Transportation routing first and cost second. However, is there a
• Porfolio analysis• better way for assessing opportunity cost?
(6) Standardization
Obviously, more DSS applications can be ad- There is a need to standardize some basic
ded to the list, and the limitations are human's features of DSSs so that they can be used by
creativity and imagination. more decision makers but at the same time
taking into account individual differences in
terms of decision style. Is there a way that the
7. Problems Facing DSSs
conflicting requirements can be resolved?
(7) Individual versus group DSSs
After a decade of building, developing and How to support individual decision makers
using DSSs, many problems have surfaced and during a group meeting is also a pressing
need to be solved eventually. We list below some problem to be resolved. Should all individuals
of the commonly seen problems: use the same DSS and see the same thing, or
should individuals be allowed to use their
(1) Data capture and collection tailor-made DSSs during a group decision
How to capture and collect data remains an making?
362 M.C. Er / Decision Support Systems
(8) Data are not independent of spreadsheets cance (King, 1984). It is an area where DSS
Data used by a spreadsheet normally cannot can make a substantial impact on the top
be used by different spreadsheet, resulting in management and the corporation. The generic
data being tied up with spreadsheets. An SICIS issue tree (King, 1984) sounds trivial,
agreeable industrial standard for all spread- and the SPIRA process for achieving informa-
sheets may not come easy. tion-based strategic comparative advantage
(King, 1984) resembles Checkland's (1981) soft
system methodology.
8. Future Trends of D S S Intelligent DSS
Some authors, notably Nolan (1986), suggest
Despite the fact that there is no acceptable the adaptation of artificial intelligence (AI)
definition, DSS is here to stay simply because it and expert systems techniques to DSS. How-
complements the transaction processing aspect of ever, most authors under-estimate the difficul-
EDP. However, it will take some time before DSS ties in representing common-sense knowledge
reaching a mature stage. As it stands, the follow- which is an unsolved problem in AI.
ing future trends of development of DSS are
identifiable:
References
(1) Group DSS
Group decision making plays a major role in Checkland, P.B. (1981), Systems Thinking, Systems Practice,
determining corporate affairs (Young, 1983). John Wiley: Chichester.
Dearden, J. (1972), MIS is a Mirage, Harvard Business Review,
How to design and develop group DSSs for
Vol. 50, No. 1, pp. 90-99.
supporting group meetings is a complex task DeSanctis, G. and Gallupe, B. (1985), Group Decision Support
because of the complex combination of peo- Systems: A New Frontier, Data Base, Vol. 16, No. 2, pp.
ple, places, time communication networks, in- 3-10.
dividual preferences, and other technologies. Ginzberg, M.J., Reitman, W. and Stohr, E.A. (1982), Decision
A group meeting can be conducted at the Support Systems, North-Holland: Amsterdam.
Greenwood, W.T. (1969), Decision Theory and Information
same place, or at different places attended by Systems, South-Western Pub.: Cincinnati.
different groups of people using teleconferenc- Hogue, J.T. (1985), MIS Manager's Guide to Decision Support
ing techniques. On the other hand, a group Systems, Data Processing Management, pp. 1-8.
meeting can be conducted during a fixed period Huff, S.L. (1985), Decision Support Systems, Computer Pro-
of time, or it is just an unlimited on-going gramming Management, pp. 1-11.
Humphreys, P. (1984), Levels of Representation in Structuring
process. Group DSS is supposed to support Decision Problems, Journal of Systems Analysis, Vol. 11,
any one of the possible combinations (De- pp. 3-22.
Sanctis and Gallupse, 1985). Keen, P.G.W. (1981), Value Analysis: Justifying Decision Sup-
(2) Decision support centre port Systems, MIS Quarterly. Vol. 5, pp. 1-15.
Decision support centre is an emerging con- Keen, P.G.W. (1987), Decision Support Systems: The Next
Decade, Decision Support Systems 3, 253-265.
cept (Owen and Volpato, 1985). A decision Keen, P.G.W. and Scott Morton, M.S. (1978), Decision Sup-
support group, staffed by information systems port Systems: An Organizational Perspective, Addison-
professionals who understand the business en- Wesley: Reading, MA.
vironment, form the core of decision support King, W.R. (1983), Achieving the Potential of Decision Sup-
centre, with advanced information technology. port Systems, Journal of Business Strategy, Winter, pp.
84-91.
A decision support centre is usually located in King, W.R. (1984), Strategic Management: Decision Support
close proximity to top management so that Systems, in manuscript.
instant decision support can be provided. A Klein, H.K. and Hirschheim, R. (1985), Fundamentals Issues
decision support group will readily develop or of Decision Support Systems: A consequentialist Perspec-
modify DSSs to support top management in tive, Decision Support Systems, Vol. 1, pp. 5-24.
Landry, M., Pascot, D. and Briolat, D. (1985), Can DSS
making urgent and importantdecisions. Evolve Without Changing Our View of the Concept of
(3) Strategic DSS 'Problem'? Decision Support Systems, Vol. 1, pp. 25-36.
DSS for supporting strategic management is a Meador, C.L., Guyote, M.J. and Rosenfeld, W.L. (1986), Deci-
well recognised area of importance and signifi- sion Support Planning and Analysis: The Problems of
M.C. Er / Decision Support Systems 363
Getting Large-Scale DSS Started, MIS Quarterly, Vol. 10, Sol, H.G. ( 1 ~ ) . Processes and Tools for Decision Support,
pp. 159-177. North-Ht~!i~.md: Amsterdam~
Nutt, P.C. (1986), Decision Style and Its Impact on Managers Sprague, R.h. jr and Carlson, E.D. (1982), Building Effective
and Management, Technological Forecasting and Social Decision S~,~port Systems, Prentice-Hall: Englewood Cliffs,
Change, Vol. 29, pp. 341-366. New Jersey.
Owen, D. and Volpato, M. (1985), Focusing High Technology Sprague, R.H. jr and McNurlin, B.C. (1986), Information
on the Executive Decision-Making Process, The Australian Systems Management in Practice, Prentice-Hall: En-
Director, Vol. 15, April/May, pp. 20-24. glewood Cliffs, New Jersey.
Pillips, L.D. (1984), Decision Support for Manager, in The ¥:,::~g, L.E. (1983), A Corporate Strategy for Decision Sup-
Managerial Challenge of New Office Technology, (ed) H. port Systems, Journal of Information Systems Manage-
Otway and M. Peltu, Butterworths: London, pp. 80-98. ment, Vol. 1, No. 1, winter.