Prioritization, Complexity, and the Analytic Hierarchy Process

The preceding chapter discussed the 7 MP tools, which are invaluable in analyzing and interpreting qualitative and verbal data. One typical challenge in such an analysis is prioritizing various tasks, issues, and characteristics that are identified in affinity and tree diagrams. This may be done using prioritization matrices, which were introduced as part of the 7 MP tools. However, prioritization matrices have limitations when it comes to dealing with a high level of complexity involving multicriteria decision problems. This limitation can be obviated by using a powerful technique, the Analytic Hierarchy Process (AHP). It helps you solve complex problems involving multiple goals, objectives, criteria, competitors, and other factors.^[1] AHP can be used to address important architecture and design prioritization issues in the software development process. As you will see, AHP and its supporting software, Expert Choice (EC), can handle much higher levels of complexities than the tools introduced in Chapter 7. In addition to solutions facilitated by EC, we will also illustrate two known approximations to AHP solutions using manual calculations. Manual calculations can be used to solve relatively less-intricate problems. We therefore emphasize EC as an important companion to AHP.

Our interest in AHP goes beyond prioritization. AHP has a wide range of applications as a multiobjective decision technique in which qualitative factors are present along with quantitative factors or are dominant. AHP has an important application in Quality Function Deployment (QFD), as discussed in Chapter 11. AHP lets you structure complexity and measure and synthesize it. It uses ratio scale measures that can be meaningfully synthesized to arrive at not only a ranking of alternatives, but assign true proportions that can be used to optimally allocate resources. AHP has a variety of applications in economics, business, agriculture, engineering, social sciences, politics, and numerous other fields. In software development it can be used for decision-making in various phases of the software development process, from requirements development to design to review, test, and evaluation to maintenance and decommissioning, involving situations with multiple objectives.

AHP is essentially a theory of measurement and decision-making developed by Thomas L. Saaty when he was at the Wharton School of the University of Pennsylvania. The real value of AHP lies in its ability to combine, or synthesize, quantitative as well as qualitative considerations in an overall evaluation of alternatives. As such, it can be especially applicable when you evaluate complex system designs involving software, hardware, and humanware that can easily include hundreds of system quality indicators.^[2] AHP has emerged as a powerful technique for determining relative worth and ranking among a set of elements. You can use it to make design, evaluation, and benefit/cost or optimal resource allocation decisions throughout the software development process.