OVERVIEW
The success of an organisation depends to a large extent upon the quality of its managerial decision-making. Everyone makes a decision when choosing between two or more alternatives. An alternative is a course of action intended to solve a problem. Although few people analyse the decision-making process, all decisions follow a similar series of logical steps as shown in Fig- ure 2.1. The decision-maker must first recognise that a problem exists. The next step is to estab- lish the evaluation criteria, i.e., the necessary quantitative and qualitative information which will be used to analyse various courses of action. The third stage involves identifying various alter- natives for consideration. Having assessed every situation, decision-makers then select the al- ternative that best satisfies their evaluation criteria. Finally, this optimal choice is implemented. Business managers must choose the appropriate course of action which is most effective in attaining an organisation’s goals. The types of decisions that have to be taken vary considerably and depend upon the complexity of the problem to be solved. Usually, the level of complexity increases proportionally with the level of qualitative information required to solve the problem. Even where a problem is based almost entirely on factual, quantitative information, decision- making can be difficult. For example, the construction of a large building contains so many interdependent activities that an overwhelming amount of time could be spent on gathering factual information about the situation. Luckily, computer models are now available to assist the managerial decision-making process!
Recognise a problem Establish evaluation criteria Identify alternative courses of action Implement this choice
Select the best alternative
Evaluate alternatives
Figure 2.1 Stages in the decision-making process.
MODELLING CHARACTERISTICS
A model is a simplified representation of a real-world situation. It can be regarded as a substitute for the real system, stripping away a large amount of complexity to leave only essential, relevant details. A model is used to facilitate understanding of a real object or situation. A decision- maker is better able to evaluate the consequences of alternative actions by analysing the model’s behaviour. For example, dummies which are used in car-crash simulations, allow engineers to test and analyse the safety of new features. The advantages of a model are (i) it is easy for the decision-maker to understand (ii) it can be modified quickly and cheaply, and (iii) there is less risk when experimenting with a model than with the real system.
Typical models showing the layout of a new housing or shopping complex can be found in architects’ offices. Such models are referred to as ‘physical’ models because they are three- dimensional representations of real-world objects. They may be scaled-down versions of the real thing or, as in the case of the crash-testing dummy, they can be exact replicas. Types of models which are more suited to computers include (i) graphical models, which use lines, curves, and other symbols to produce flow charts, pie charts, bar charts, scatter diagrams, etc. and (ii) mathematical models which use formulae and algorithms to represent real-world situations. All of the models developed in this book are either graphical or mathematical.
Because a model is not an exact representation, it cannot contain all aspects of a problem. Models involve a large number of assumptions about the environment in which the system operates, about the operating characteristics of its functions, and the way people behave. These environmental assumptions are called uncontrollable variables because they are outside the decision-maker’s control, e.g., interest rates, consumer trends, currency fluctuations, etc. On the other hand, controllable variables are those inputs that influence a model’s outcome and are within the decision-maker’s control. Examples of control variables are product price, the level of output, or acceptable profit levels.
Steps in Model-Building
The validity of a model’s results will depend on how accurately the model represents the real situation. The ideal model is one which is neither too trivial in its representation of reality nor too complex to implement. There are three main steps in building a spreadsheet model as shown in Figure 2.2. The first step of problem formulation requires that the decision-maker adopt a systematic and rigorous approach to the selection of variables. One of the main sources of errors in model-building is the exclusion of important variables, due either to an oversight or a lack of understanding of the model’s underlying logic. In either case, the results will be meaningless or at best, dubious.
The second step in model-building is the identification of the relationships among the vari- ables. In many cases well-known formulae may already exist. For example, consider the amount due to an investor at the end of five years when £2000 is invested at a compound interest rate of 12% per annum. The general formula for basic compound interest states that if an amount P0is invested at a fixed rate of r %, the principal Pnafter n years is
Formulate the problem Identify the relevant variables
Distinguish between controllable and uncontrollable variables
Build the model
Enter data and formulae into the relevant spreadsheet cells Establish mathematical formulae
Define the relationships among the variables, i.e., the model’s formulae
Figure 2.2 Steps in model-building.
The investor will thus get 2000(1.12)5 = £3525. In this case, the controllable variables are
P0, r , and n. In situations where no known formula exists, the decision-maker must take great
care in defining the relationships among the variables. A sensible precaution is to check out a newly-derived formula by using historical data and comparing the formula’s result with an already-known answer.
The final step is to set up the model as a spreadsheet. The compound-interest example above is used as a demonstration model. A spreadsheet is usually depicted as a matrix or grid of cells, each uniquely numbered by reference to the column and row in which it is found. A cell can contain either a description, a fixed value, or a formula, e.g., cells A1, E3, and E7 in Figure 2.3 show each of these three features. The main source of errors in building a spreadsheet model is due to wrong cell cross-referencing, i.e., the user incorrectly identifies source cells when entering either input or formulae.
The advantage of the spreadsheet model is that the user can easily change any of the input data. Thus the compound interest for any values of P0, r , or n can be found by simply overwriting
the current input values in cells E3, E4, and E5. The model also allows users to see instantly the
A
1 Compound interest example
2
3 Input data Initial principal, P0
Principal Pn after n years is £3,524.68
4 Interest rate, r 5 No. of years, n 6 7 8 9 B C D E F G H I £2,00 12% 5
Cell E7 contains the formula E3*(1 + E4) ^ E5 where the symbols *, ^ represent multiplication and
' to the power of '
Table 2.1 Characteristics of the decision-making environment.
← The decision-making environment →
Characteristics Certainty Risk Uncertainty
Controllable variables Known Known Known
Uncontrollable variables Known Probabilistic Unknown
Type of model Deterministic Probabilistic Nonprobabilistic
Type of decision Best Informed Uncertain
Information type Quantitative Quantitative and qualitative Qualitative
Mathematical tools Linear Statistical methods; Decision analysis;
programming Simulation Simulation
effects of their deliberations. By experimenting with various combinations of variables, the decision-maker can obtain a better understanding of the model’s sensitivity.
RISK AND UNCERTAINTY IN DECISION-MAKING
In business decision-making, there are three classes of decision problems, namely