The Utilization Subsystem: Decision Making-Problem Solving

The three fishermen decided to fish for the day. At the camp, they arranged to rent a small motorboat for their outing. The agent at the camp, who arranged for the rental of the boat, alerted the three fishermen that they would have to return the boat no later than 5 p.m. that evening, just before sundown. He could not insure their welfare beyond that point in time. He also informed them that the main lake flowed into three adjoining lakes and that the second lake represented a better source for catching the big fish. The three fishermen boarded the motor craft, hoping for an enjoyable day fishing. The three explored the first lake without success. They explored the second lake with a similar outcome. While on the third lake, they were so successful that the time of day escaped them. At one point in time, one of the fishermen informed the others that it was late and that they should return so as to obey the time deadline presented by the agent earlier that morning. While returning, a major problem surfaced. There were so many outlets to the three lakes that it was impossible to establish which outlet they had left that morning. Without a map or another source of information, they were unable to decide which inlet would lead them to the camp safely. In addition and unfortunately, there were no other craft on any of the three lakes that could be summoned to provide them with the information they required to help them to return safely to the camp. Fortunately, the camp who had leased them the motorboat had an adequate information system to locate them and return them to camp safely [true story].

LEARNING OBJECTIVES



OVERVIEW


Each day, individuals are faced with making decisions. Ways of thinking, as well as conduct, can influence the quality of our lives and the objectives and goals that we set to achieve. Human decision making has challenged many minds throughout our history. The arrival of the computer has stirred even greater interest for many scientists attempting to identify the properties of human decision making. In fact, an augmented data, information, and knowledge (ADIK) system's most important role is to aid the decision process. It extends our native cognitive ca• pacities. This chapter describes those aspects of interest to the information scientist, whose challenges are extensive and sometimes difficult to face. First, the critical terms and ideas related to decision making are defined. The focus of attention is our current understanding of the decision process, including problemsolving and decision analysis and the computer tools that aid these processes. Included is a discussion of artificial intelligence and expert systems and how these relate to the human decision function.


The Study of Decision Making and Problem Solving


All of us are familiar with the words "problem" and "decision." A "problem" is something difficult to deal with and understand. More often than not, it involves decisions to be made, decisions that will affect the solution of the problem. We make decisions in almost all the things we do, from going to scbool, to the baseball team we watch, to the kind of work we want to do. Decisions require making choices. Decisions often do not come alone. They are generally part of problems we face. A problem is related to some goal, such as the best place to go to college and support oneself in the process. Problems are obstacles; decisions involve finding ways of facing and overcoming these obstacles. There may be no single, best way to achieve the end goal. There are countless conditions and states we face each day. Each involves decisions to be made. Even postponing a decision could have important consequences. Each of these decisions demands data, information, and knowledge to support actions to be taken. The possibility that terrorists would apply biochemical warfare to achieve their goals raises many problems and may require many decisions to be made. The events of September 11 present a horrendous number of problems and with each, an important decision must be made. No other area in information science merits more careful and direct atten tion than human problem solving and decision making. Each of the segments of the knowledge sciences-documentation, computerization, and communication-includes the human action of problem solving and decision making. Humans have an enormous capacity to solve problems and make decisions. We can see this all around us. Particularly, think of the many advances that have been made in transportation, medicine, and other fields based on the advances in physics, chemistry, astronomy, and the other sciences. Much of our capacity for solving problems and making decisions is based on the capacity of the human brain. However, the technology that humans have invented and created extends this capacity. Whether it is in private or public life, the human requirement to augment human capacity to make decisions and solve problems is paramount. Decision making and choice go hand in hand. Humans have a significant capacity for making choices, particularly in certain societies. The greater the choice base, the greater the requirement for aids that help or assist in making choices and resolving problems. One can say that a core issue in information science is understanding how technology aids individuals in meeting conditions that require actions, whether these actions are to resolve conflicts, take advantage of opportunity, or safeguard personal welfare.


Definitions


The Problem


A problem is when there is a difference in the actual state of a situation and the desired state. A person is confronted with a problem when he wants something and does not know immediately what series of actions he can perform to get at it. The desired object may be very tangible (an apple to eat) or abstract (an elegant proof of a theorem). It may be specific (that particular apple over there) or quite general (something to appease hunger). It may be a specific physical object (an apple) or a set of symbols (the proof of a theorem). The actions involved in obtaining desired objects include the physical actions (walking, reaching, writing), perceptual activities (looking, listening), and purely mental activities (judging the similarities of two symbols, remembering a scene, and so on) (Newell and Simon 1972, 72).


Problem Solving


What is problem solving? Problem solving is "cognitive processing directed at transforming a given situation into a goal situation when no obvious solution method is available to the problem-solver" (Mayer 2002, 37).


Problem Space


A problem space is a set of elements expressed in symbols representing a state of knowledge; operators that change an existing state of knowledge; an initial state of knowledge; a problem representing a desired state of knowledge and prevailing knowledge in the area of interest to the person faced with the problem (adapted from Newell and Simon 1972, 810). The important point to information scientists, as we have learned, is that all humans are limited in our innate capacity to solve problems and make decisions. The challenge is to determine the best ways to use the tools that we have made possible to help us create, invent, or revise the ways to do things better, faster, or cheaper when the need and requirement arise. That is the function of an AD IK system.


Decision Making

John Dellen had to make a decision: go out for the basketball team or work for a 3.5 grade point average to get into his college of choice. In this case, being on the basketball team and obtaining good grades depend on each other. Both require equal amounts of effort. Having to work after school so that he could enjoy the few hours he had with his friends came into the picture as well. How to manage all these options was a problem for John. What could be done about all this? Perhaps a little thinking on his approach might help. Decision analysis is such an approach.


Decision-Making Analysis

The decision refers to making a choice among different alternatives, such as going to school or going to work to earn a salary to support a family. Analysis means finding out similarities and differences about states, conditions, aspects of problems, etc. In a word, decision analysis is understanding a situation. It includes estimating risks and minimizing losses. It refers to deconstructing or breaking down problems; understanding their structures; determining their inherent values; and developing a model of all of the various aspects of the problem and the decision required. The key to decision analysis is reducing a problem to its basic element. Some decisions are easy to make; others are more difficult. It depends on the problems one faces and the consequences of the outcomes. We should realize, at the outset, that almost all of our actions involve decisions. Now and then we have problems to face, such as how to study for tomorrow's exam and still watch television tonight. Or how about what we should do after graduation: go to college or get a job? Your mom bought a new VCR to watch and record your favorite TV shows, but its manual is really unclear on how to set up the recording. The problem: how to operate the new device. The decision to be made: should you bother your mom about it or just not use it? Which decision would be wise? Problem solving and decision making are part of everyday life. Decision making and problem solving have been studied by behavioral scientists, economists, management scientists, and others for many years and much has been learned about decisions and how we make them. The information scientist uses this knowledge when he or she analyzes and designs ADIK systems to deal with the many events and situations we are presently faced with and those that we are expected to face in the future. One of the basic aspects of problem solving and decision making is that we can develop tools that help us solve problems and make decisions. There are so many tools that we might mention. It would be almost impossible to list them all. The ordinary pencil and pen, for example, are used each day to help us solve problems. We use symbols to create all sorts of things including numbers, letters, and pictures, which we put together on paper or in our heads in an attempt to come up with an answer or approach to a problem. We use the old-fashioned barometer outdoors to tell us whether we should wear a heavy coat or take an umbrella to school on a particular day. We use a car speedometer to help us make a decision as to when to speed up and pass a car within a certain speed limit zone; the pilot watches the altimeter to establish how high an aircraft is above the ocean or a mountain. We can go on and on giving examples of technologies that we have invented to help us make decisions and solve problems. So where and how does our study of decision making and problem solving fit in our study of information science? If one was to ask what tool can be considered to be most important in helping us solve problems and make decisions, we most certainly would consider the computer as the obvious choice. The computer, however, is not the only choice. If we take into account the data, information, and knowledge systems that we study, we would include all the technologies we mentioned above plus, most importantly, such technologies as sensors, eyeglasses, binoculars, cameras, radar, sonar, telephones, satel· lites, and much more. There are other candidates such as artificial intelligence, expert systems that information scientists include, that we will look into later in this chapter.


Software Tools and Decision Analysis


We know that software enables humans to use computers for their own purposes (applications). There are many software programs that can be used for a variety of decision analysis applications. Information and computer scientists are continually developing new software to support human problemsolving and decision-making functions. There is a chain of logic that helps us structure a decision, that is, to see how various parts of the required action follow from each other in achieving an objective. There are software programs that help people make a decision while taking into account the value of the information that comes into play when a decision is to be made. There are software packages that provide us with the options that are available in dealing with a problem or decision situation, given the information at hand, and the best way to act (Clemen and Reilly 2001, 10). Although decision analysis provides a good way of looking at how we make our choices-while providing ways of looking at alternatives-it should not induce us to accept the alternatives that it provides. As a matter of fact, if good decision analysis is engaged, there should be no need for blind acceptance of decision alternatives. It should provide a better understanding of the problem rather than provide solutions. Decision analysis is an information source to clarify objectives, alternatives, and trade-offs. Decision analysis does not take over the decision maker's job. It helps do the job. Decision analysis helps us make inferences about what is happening or is likely to happen. It offers possible alternatives to action and a trade-off for each, thus helping us negotiate appropriate actions to be taken. Decision analysis allows for personal judgment to enter into the decision process. In fact, it requires it. Ta ble 8.1 shows an example of decision analysis modeling. There are certain aspects of a decision that relate to its difficulty. The decision may be hard because it involves numerous factors, some related and others not. Decision analysis can help group these factors together, examine each carefully, and evaluate similarities and differences before taking action. Another difficulty with decision making is the uncertainty of outcomes. There may be many objectives desired, all with different costs and values to each (see figure 8.1). Passing an exam may get one into college, but on the other hand, one may not be able to gain a scholarship based on athletic competition. So there may be a question of the price for achieving one objective at the cost of another. It can come down to the matter of having to see a problem in different ways. For example, a mom and a dad could have different views of the achievements of their son. Again, the decision may be difficult if the father is factored into the decision process opposite the mother. So how can anyone establish the bestdecision for any particular situation? Figure 8.1 shows how different types of decisions are made by different managers. Decision research tells us that a good decision is one that provides the "best" attainable outcome. We can say, "Of course:' But then, where does luck come into the picture when decision analysis is applied? Decision researchers tell us that decision analysis cannot improve your luck. Yet it can improve your understanding of how to make better decisions. This understanding includes the structure of the problem and how to deal with the uncertainty related to it. Decision analysis helps those who are sufficiently intelligent and thoughtful to look at the problem and the decision carefully and then establish or prescribe the action that could lead to the best results. Although we are not perfect decision makers, we could do better by looking at a problem carefully and systematically. Information science is a major player applied to this end. Matching the human with the technology could do the trick. With this background, we can now examine the various programs computer scientists have provided that can be applied directly to our study of decision analysis. In this process, we will also examine how artificial intelligence and expert systems complement decision-making and problem-solving functions, acknowledging that each could be a study in its own right.



Computer Tools That Aid the Problem-Solving and Decision Process


Computer programs have been designed to help the modeling and the solution phase of the decision process. All of these programs work together. One example is Microsoft Excel. Excel is a program that employs an ordinary spreadsheet, that is, a working sheet consisting of columns and rows that help in viewing and working with data. Such a tool is best used to model uncertainty. Uncertainty is a major human cognitive (mental) state related to information which enters in many decision-making activities. These may include managing the research and development of projects and programs; determining the best way to discover and use resources (human, oil, energy, etc.); bringing new products to market; or how best to deal with disaster (9/1 1 , hurricanes, tornadoes, medical epidemics, etc.).



Artificial Intelligence/ Expert Systems

Artificial intelligence started with the idea that computers can really help us think and learn. How can computers help us solve problems and make decisions? How can ADIK systems help us solve problems and make decisions? Let us first define terms and then show how information scientists and others are applying and conducting research to understand the best way to aid decision making and problem solving in specific applications. Artificial intelligence (Al) has been defined as "the ability of a computer to 'think' for itself. AI studies usually focus on understanding how humans think and how these capabilities might be instilled in the computer which includes speech recognition, deductive reasoning capabilities, creativity, and the ability to learn from experience (as opposed to memorizing data)" (Computing Dictionary 1998, 86). Expert systems are considered to be a form of artificial intelligence, "an application that makes decisions by using facts, rules, and a reasoning ability called an inference engine. The facts are supplied by human experts in a particular field. Common categories include medicine, investments, automobile routing, insurance, equipment repair, and science" ( Computing Dictionary 1999, 154). Expert systems have, at times, been referred to as "knowledge systems" when applied to practical problems that information professionals face. "Knowledge systems in general do not necessarily mimic human experts, but do provide the electronic means to collect, store, distribute, reason about, and apply knowledge. Expert systems are a specialized case, incorporating knowhow gathered from experts and designed to perform perishable human expertise; distribute otherwise scarce expertise; reduce costs of Medicare or poor human performance; provide help to humans trying to access information and use computers" (Smith 1987, 51). At the outset, it is understood that almost all activities engaged in by information professionals include a large measure of problem solving and decision making. These functions include the access to and acquisition of data, information, and knowledge, the processing of these resources, and the actions ( decisions) demanded by the task. The following are some of the current interests of information scientists in applying artificial intelligence and expert systems to problem solving and decision making:


  1. What are the means for updating data automatically lo meet needs and requirements of users?
  2. How can one organize knowledge in such a way as to aid us to think, learn, and use the literature efficiently and effectively (Heilprin 1989; Hjerppe 1992)?
  3. How can one establish the best interface between the human and the computer in getting things done?
  4. How can one identify ways that the computer can help human memory in processing and updating of information and knowledge?
  5. How is it best to apply expert systems in reference work in libraries?
  6. How does one codify and improve what we know when it is broken up, full of mistakes, and make what we understand more precise and reliable?
  7. How does one determine the social impacts of AI and expert systems on decision makers and problem solvers in our present institutions and environment?

Summary


Most of our lives involve problems that we have to confront and decisions that we must face before we act. Data, information, and knowledge systems are analyzed, designed, and evaluated to serve this purpose. On this basis, we can act prudently, knowing how the information systems we create truly augment, that is, extend our capabilities to perform these two important human functions. Decision making and problem solving consume much of the capacity of our brains. They provide the experiences we accumulate in the course of dealing with life's circumstances. Analysis of problems and decisions related to them is a matter of historical reference. The approaches to them, however, have changed over time as insight and technology have become more available to deal with them. One tool, among many, is decision analysis. This method of dealing with decision making includes procedures such as detailed identification of the variables (factors) of the problem faced, modeling of the problem space, the sequential tracing of steps in the problem, a solution phase, identifying alternative options for problem solution, and determination of the risks and costs related to them. Toward this end, information and computer scientists have provided a variety of software programs that aid these processes. Artificial intelligence and expert systems represent a class of such tools. They seek to have computers help us think, learn, and hopefully act. Information scientists are continuing to study how these resources can be applied to help us deal with the many complexities of life we face.


EXERCISES

  1. You have been admitted to three colleges to which you have applied. Each has merits that can be easily identified. The selection of one, however, could be difficult due to the fact that they have equal advantages to you. Identify and discuss those factors that are included in this problem decision. How could an information professional help you make a decision?
  2. What is the factor that is considered most important in analyzing a problem or making a decision?
  3. Discuss why decision making and problem solving should not be solely centered on computer technology.
  4. Select a problem of your preference and show how modeling can help in the solution of the problem.
  5. Identify how information professionals apply an understanding of decision making and problem solving to library operations.
  6. How does decision theory relate to information retrieval?
  7. What is the relationship between a knowledge system and artificial intelligence?
  8. How and when might an expert system fail to support decision making and problem solving?
  9. Describe a current problem you face. Sketch a model of such a problem that could serve as the basis for decision analysis.
  10. Most, if not all, problems relate directly to an objective. Discuss how the satisfaction of an objective can be evaluated.