Technological Forecasting
|
Judgment-Based Technological Forecasting Techniques
Network Analysis Network analysis is a formalization and extension of monitoring. This technique can be used in two distinct ways: (1) as a method of exploring the possible capabilities and systems that might result from extensions of current scientific research and, (2) given a desired capability or end system, as a method for determining what research results are required to achieve the desired capability. The first use is exploratory forecasting, and the second is prescriptive forecasting. In the former, possible future technologies are explored; in the latter, desired future capabilities are defined. The term network model results from the fact that networks are commonly used to organize the data. Real examples of the method are quite complex, so consider a hypothetical example. Assume we have been monitoring the appropriate sources, and we have knowledge of recent research results, rl, r2, . . . rk. We predict that these might be combined as shown in Figure 10, to yield s1, s2, . . . sj scientific capabilities. These capabilities, in turn, could be combined to result in t1, t2, . . . tm technical components. Finally, the technical components can be joined to produce end systems, V1, V2, , . . .Vn (Figure 11). 
Going from research results to end systems is exploratory, but the direction may be reversed. instead of seeking the system implications of research results, we could begin with one or more related desirable end systems. Going backward through the same chain of logic, we seek the components, and the research results necessary to yield the capabilities. This uses the network to prescribe research. If the various end systems are given weights that reflect their relative importance, it becomes possible to set priorities on the various research results based on the importance of the end system(s) to which the research contributes. One method of finding these priorities is described by Dean (Chapter 11 in [6]) and discussed in detail in Chapter 2 of this book. Dean developed the technique to evaluate research labs after the fact, but if the "evaluation" is a priori, it becomes a statement of priorities. It follows that if we know the cost of the research projects, can estimate the relationship between research results and research expenditures, and have a budget constraint, mathematical programming (see Chapter 9) can be used to determine the best order and extent to fund the various research projects. It is also possible to treat both the prescriptive and exploratory models stochastically--that is, to estimate the probability that the research/capability, capability/component, and component/system connections can be achieved. For the reasons explained in Chapter 2, we do not see this practice as significantly helpful. In any event, no matter how many numbers are applied to the network, the base information is strictly judgmental. [Back to Table of Contents] [Next Page] |