管理信息系统（第7版）PPT+试题库 Ch10-7E.docx

Essentials of Business Information Systems, 7ELaudon & LaudonLecture Files by Barbara J. EllestadChapter 10 Improving Decision Making and Managing Knowledge"Companies have been able to use technology to do some very cool stuff-to reach customers in new ways, to automate operations. But one thing many businesses haven't been able to do easily is use the data they've collected to find and stamp out waste across operations. Sifting through corporate data was supposed to make executives more efficient. Much of the time, though, it's just made them more confused/9 (Fortune magazine, March 3, 2002)10.1 Decision Making and Information SystemsEach of us makes hundreds of decisions every day. If just a fraction of those decisions could be improved through better and more information and better processes, we'd all be delighted. Businesses feel the same way. Customers would be happier, employees would be more motivated, and managers would have an easier job. Most of all businesses could improve their profitability to the benefit of all.Business Value of Improved Decision MakingTable 10.1 provides a few examples of the dollar value that enhanced decision making would give to firms.Production: Insert Table 10-1Don't be misled into thinking that the dollar value of improved decision-making process is limited to managers. As more business flatten their organizational structures and push decision making to lower levels, better decisions at all levels can lead to increased business value.Types of DecisionsAnd for geographically separated attendees, travel time and dollars are saved. Electronic meeting systems make these efficiencies possible.Group InteractionBrainstorming Topic Commenter Group OutlinerIdea GenerationIdeaOrganizationIdea OrganizerIssue AnalyzerGroup riterPrioritizingOrganizational MemorySessionManagerPolicy ormation Sta eholder IDo te Selection Alternative Eval uestionnaire Group MatriPolicy DevelopmentSession Planning Personal productivityFigure 10-5: Group System Tools.Enterprise AnalyzerGraphical BrowserGroup DictionaryBrief Case Access to informationFigure 10-5 shows a typical meeting using group system tools.All is not perfect with GDSS, however. Face-to-face communication is critical for managers and others to gain insight into how people feel about ideas and topics. Body language can often speak louder than words. Some people still may not contribute freely because they know that all input is stored on the file server, even though it is anonymous. And the system itself imposes disciplines on the group that they may not like.Bottom Line: Executive support systems meet the needs of corporate executives by providing them with vast amounts of information quickly and in graphic form to help them make effective decisions. ESS must be flexible, easy-to-use, and contain both internal and external sources of information. Using group decision-support systems, comprised of hardware, software, and people, helps streamline group meetings and communications by removing obstacles and using technology to increase the effectiveness of the decisions.10.3 Intelligent Systems for Decision SupportMany people have the impression that artificial intelligence (AI) is all about computers taking over the world and turning on their human inventors. That's not true; they can't replace humans. Many of the systems under the AI umbrella are useful tools for capturing, storing, and disseminating human knowledge and intelligence.Expert SystemsExpert systems are a common form of artificial intelligence. They are used to assist humans in the decision-making process, but they don't replace humans. Many of the decisions we make are based on past experience, but we have the added benefit of reasoning and intuition. Expert systems ask questions, then give you advice and reasons why you should take a certain course of action based on hard data, not on hunches.Again, they don't make the final decision.Most of the problems an expert system helps resolve can, in fact, be solved by a human. But since the computer is faster or safer, businesses choose to use them instead of a person.How Expert Systems WorkExpert systems rely on a knowledge base built by humans based on their experiences and knowledge. The base requires rules and knowledge frames in which it can process data. When you think about it, humans work the same way. You look out the window to see if it's raining, /fit is, then you grab your umbrella, /fit's not raining, then you don't. There you have it, a rule base.Yes, we used a very simplified example. Most expert systems require thousands of rules and frames in which to operate in a rule-based expert system. The knowledge must be specific. In the example above, you wouldn't take any action if the only information you had was "It rains 350 days a year in the Amazon rain forest.Neither would an expert system.The AI shell (the programming environment of an expert system) uses rules, frames, and an inference engine to accomplish its tasks. The inference engine moves through the rules and frames until it finds an appropriate one and then uses it.A knowledge engineer is especially adept at pulling information from various sources, including humans, and making sure it fits into the expert system. The hardest part of the job may be convincing people to offer up their expertise and knowledge that can be incorporated in the system.Examples of Expert SystemsYou measure the success of an expert system by the following: Reduced errors Reduced cost, reduced training time Improved decisions Improved quality and services Happy users and happy customersMost problems solved by expert systems are mundane situations. "If it's raining, then take an umbrella." But what happens if it's cloudy and only looks like it will rain? Expert systems only do well in situations in which there are definitive outcomes. They aren't good at making decisions based on inferences. The expert system might advise to take the umbrella along or to leave it home based on the input. The human makes the final decision to take or leave the umbrella.If you understand that expert systems can only do so much, you'll be just fine. If you understand that they aren't people with the powers of reasoning and intuition, and therefore they can't make every decision, you'll know when to override the system and when to go with its output. Remember that everything in an expert system is based on IF this, THEN that. We know not everything is black and white and there are many gray areas.Case-Based ReasoningSo far, we've concentrated on capturing individual knowledge in an expert system. Through practical experience, you've realized that "two heads are better than one.” Very seldom will only one individual work on a project. Or perhaps one individual works on the candy bar ad campaign while another works on the breakfast cereal campaign. They have different and yet similar experiences. What if you could tap into each person's experience and knowledge on a collective basis? Take the best of the best from each one and apply it to your needs. Then you give your knowledge to someoneelse who will combine it with knowledge from others and continue building on “the best of the best." That's what a case-based reasoning (CBR) system does best.The Help files you find in most desktop software applications are built on a case-based reasoning model. The technical support staff combines thousands of customer queries into a single database of problems and solutions and refines that information into a series of IF this is the problem, THEN try this. Access the Help files in your desktop software and try it.5.2.3.4.6.User describes the problemSystem finds closest fit and retrieves solutionSystem stores problem and successful solution in the databaseSystem searches database for similarcasesCasedatabaseSystem asks useradditional questionsto narrow searchSystem modifies thesolution to better fitrthe problem<Successful?NOYESFigure 10-7: How Case-based Reasoning Works.Figure 10-7 gives you an excellent overview of how a case-based reasoning system works.Fuzzy Logic SystemsOkay, one more time, back to our umbrella. If it's only cloudy outside, how do you know whether to take the umbrella? ”It depends on how cloudy it is,“ you say. If looks like rain, you know to take the umbrella; there is a strong possibility that it will pour buckets. If it's only a little cloudy and doesn't look like rain, you'll take the chance that you won't get wet and leave the umbrella at home. That's fuzzy logic!Fuzzy logic, a relatively new rule-based advance in AI, is based on approximate values and ambiguous data. A fuzzy logic system will combine various data into a range of possibilities and then help solve problems that we couldn't solve before with computers.Neural NetworksThis type of knowledge system is as close to emulating the human ability to learn as we've been able to come. Let's return to our umbrella example. How do you know to take an umbrella when it's raining? You probably got wet a few times without one. Then you tried using one when it rained and discovered that you didn't get wet. You learned that when it rains, an umbrella will keep you dry. That's basically how neural networks work.You give a neural network data for which you already know the output, so that it has a base of correct information upon which it can build. When you give it new, different data, the computer will compare it with the previous data to determine what the correct outcome of the situation should be. If the data don't fit, it figures out why. It adds that information to its current database of knowledge and then keeps taking in more data. It eventually learns the right outcome. The more data it takes in, and the more situations it gets right, the better it becomes at knowing the right answer to the next set of decisions.Figure 10-9 shows how a neural network operates.Input LayerHidden LayerOutput LayerResultsValid purchaseFraudulent purchaseData Age Income Purchase history Frequency of purchases Average purchase sizeFigure 10-9: How a Neural Network Works.The Difference Between Neural Networks and Expert Systems Expert systems emulate human decision making. Neural networks learn human thought processes and reasoning patterns. Expert systems use rules and frames in which to make their decisions. Neural networks adjust to inputs and outputs. Expert systems require humans to update their database of information. Neural networks continue to expand their own base of information.Genetic AlgorithmsWe've evolved as a human race through genetics. We are made up of many combinations of generations of humans. That's how genetic algorithms work. Solutions to problems are examined by the system. The best solution is retained for future use, while the worst solutions are discarded. The solutions that are retained are used to help provide better solutions to future problems. They are combined and changed the next time they are used.Businesses often need to solve problems that are dynamic, complex, and have many variables. Very few problems are clear-cut, black-and-white. Genetic algorithms are good systems for businesses to use because it's almost like having millions of people coming at a problem from all directions.Intelligent AgentsJump on the Web and find the best price for computer printer supplies. Simply typing the words "computer printer supplies“ into your favorite search engine will result in thousands of pages with more than just price information. You can find specific information on prices much faster using an intelligent agent. These software programs learn your personal preferences for accomplishing simple tasks and can take the drudgery out of repetitive, specific work. Figure 10-11 in the text demonstrates intelligent agent technology at work.Businesses can use intelligent agents to help train users on new systems, schedule appointments, or monitor work in progress. By far though, the most popular use of this nifty little software program is as a "shopping agent” that surfs the Web for you looking for specific items to purchase or the lowest prices on a particular item.If you'd like to try a shopping bot yourself, try <A HREF="http:” target=,new,>MySimon</a>. The Web site explains its service this way “Our secret is a team of helpers built with patent-pending software. The Virtual Learning Agent technology creates "intelligent agents5 trained by our own team of shopping experts to collect information from any online store." It's fun and fast.Bottom Line: Businesses are interested in artifleial intelligence to preserve the intelligence and knowledge of their employees and use it to their competitive advantage. Expert systems emulate humans in the decision-making process but cannot replicate the intuition and reasoning that still require the human touch. Many new technologies can help humans solve difficult problems or take advantage of new opportunities. Neural networks learn how to make decisions. Fuzzy logic uses ranges of possibilities instead of giving black-and-white, yes-no answers. Intelligent agents take much of the drudgery out of searching dozens of Web sites.10.4 Systems for Managing KnowledgeCreating and using knowledge is not limited to information-based companies; it is necessary for all organizations, regardless of industry sector. It's not enough to make good products. Companies must make products that are better, less expensive to produce, and more desirable than those of competitors Using corporate and individual knowledge assets wisely will help companies do that.All the way back in Chapter 1 we discussed the difference between computer literacy and information literacy. We pointed out that there is more to information than just bits and bytes. The next step up from information literacy is knowledge. An organization must transform the information it gathers and put it into meaningful concepts that give it insight into ways of improving the environment for its employees, suppliers, and customers.Enterprise- Wide Knowledge Management SystemsKnowledge exists throughout the enterprise in three basic forms: Structured text documents such as reports and PowerPoint presentations Semistructured such as emails, brochures, pictures and graphics Tacit knowledge that resides inside people's minds and is mostly undocumentedThe goal of enterprise-wide knowledge management systems is to capture as much as possible of these three kinds of knowledge. Once it's captured, the system must provide an easy way