Chapter 13: End-of-Chapter Questions

1. Simon distinguished between programmed and nonprogrammed decisions, and identified a four-step process for solving problems.

2. Gorry and Scott Morton initially viewed the DSS as a category of applications yet to be successfully performed on the computer.

3. Alter’s first three DSS types (Retrieve information elements, Analyze entire files, and Prepare reports from multiple files) can refer to database queries. These outputs are special reports. The two types that deal with files can also apply to the preparation of periodic reports by means of custom software or report writers.

4. If expert systems fit into Alter’s taxonomy they would be classified as Make decisions. There is strong support for this inclusion. Peter Keen, in an address to a DSS conference in the early 1980s stated that the DSS concept must embrace artificial intelligence in order to survive.

5. None of the CBIS concepts embrace the idea of replacing the manager. They all assume that the manager will be involved, not only in developing the systems, but in using the output.

6. The GDSS supports problem solving by keeping problem solvers focused on the problem, and allowing more time for identification and evaluation of alternatives.

7. The four GDSS environmental settings are the decision rooms, local area decision network, legislative session, and computer-mediated conference. The legislative session can handle the most participants, and the decision room the least.

8. A telephone call is a synchronous communication; e-mail is asynchronous.

9. In parallel communication, all GDSS participants can send and receive information at the same time. The feature that makes it impossible to link a particular participant with a particular comment is called anonymity.

10. Lotus Notes captured the groupware market because of the backing by IBM and Lotus. Also a factor was its ability to handle the communications traffic of very large firms.

11. The key DSS word is decision; the key GDSS word is communication.

12. Benchmarking enables the firm to make an apples-to-apples comparison of vendor products and to make certain that the products will solve the firm’s problem. Vendors participate when the size of the potential order is large.

13. The term consultation is appropriate since the expert system plays the role of a consultant in advising the user concerning a problem solution. The term recognizes that the user has the final say in making the decision.

14. The manager enters instructions and information into the expert system.

15. An expert system produces explanations of questions that the user asks, as well as the problem solution that is reached by the expert system.

16. The name goal variable is used to describe the solution that is sought. There can be more than one goal variable. Expert systems that classify phenomena have multiple goal variables—one for each phenomenon.

17. A rule is only fired when it is true, not when it is false or the status is unknown.

18. The forward reasoning process is concluded when no rules are fired on a pass through the rule set. Students invariably ask why the process is not stopped when a value is assigned to the goal variable. The answer, which applies only to forward reasoning, is that subsequent passes might produce one or more better solutions.

19. Reverse reasoning is faster than forward reasoning because all rules in the rule set do not have to be examined.

20. Forward reasoning is preferred when there are multiple goal variables, all or most all of the rules in the rule set must be considered, and the number of rules is relatively small.

21. The two primary approaches to developing expert system software are programming languages and expert system shells.

22. Expert systems are limited in handling inconsistent knowledge and in applying judgment and intuition.

23. Hebb’s learning law recognized that the effect of one neuron on another is related to the frequency of the output of the first on the input of the second. In other words, when one neuron continually bombards a second neuron, the amount of influence increases. Students should be able to relate to this in terms of how they have learned from their parents, teachers, friends, and so on. For example, after your parents continually tell you to “Eat your broccoli,” you finally get the message.

24. Rosenblatt’s perceptron could not handle exclusive OR logic.

25. An ANS provides a learning ability, whereas a conventional expert system does not.

Topics for Discussion

1. Gony and Scott Morton felt that management level contributed the most to decision support since that is the axis of the grid that influenced the degree of success in applying the computer.

2. It’s easy to confuse programmed/nonprogrammed decisions and structured/unstructured problems. To the best of the author’s knowledge, nobody has clearly addressed the connection. One would assume, however, that the manager makes nonprogrammed decisions when solving an unstructured problem and makes programmed decisions when solving a structured problem.

When solving a semistructured problem, the manager would make programmed decisions to solve the structured part and make nonprogrammed decisions to solve the unstructured part.

3. Neural networks and their incorporation of learning capabilities can benefit business by providing machines that can improve performance over time as experience is gained with a process. Perceptive systems offer business the potential of machines that can use their sensory perceptors and function in a way that is similar to human workers. Robotics offers the opportunity to reduce manufacturing costs and improve the quality of some manufacturing processes through the use of factory robots. AI hardware can be incorporated into machines to achieve certain levels of intelligence and also sensory capability. Natural language processing can benefit business in the area of human-computer interaction by providing the basis for revolutionary languages of the future.

4. An expert could achieve consistency and perhaps speed through the use of an expert system that she or he helped to develop that would otherwise be impossible.

5. Executives shoulder the blame for expert system failures due to changed focus of the firm, solution of a non-mission critical problem, and exposure to legal liability. Information specialists are responsible for high system maintenance cost and incompatible systems. Users are to blame for resisting a system developed by outsiders and refusal to assume maintenance responsibility. Information specialists and users jointly are to blame for changed tasks and underestimated size of the task. Loss of key development personnel is the fault of the management team in general—on all levels and in the areas where the employees worked.

Topics for Discussion

1. The reduced facility and equipment costs are most easily measured in monetary terms since that data resides in the AIS database. A review of personnel records could reveal whether a larger number of persons who were previously unemployable are now on the company rolls. Likewise, a review of historical data, such as that maintained by the labor relations office, could reveal if the virtual office has any impact on work stoppages. The advantages of the formal communication network would have to be measured subjectively by persons working in the system.

2. Firms are most likely to take an organizational approach to implementing E-mail, voice mail, audio conferencing, video conferencing, computer conferencing (since it uses E-mail software), and imaging. Individual users are most likely to implement word processing and electronic calendaring. Either firms or individuals can implement FAX, videotex, or DTP.

3. The only applications that must be used in a synchronous fashion are audio and video conferencing. E-mail can be used in either synchronous or asynchronous modes. All others are asynchronous.