Step 4: Explore Systematically
As a result of the external and internal search activities, the team will have collected tens or hundreds of concept.fi'agments- solutions to the subproblems. Systematic exploration
is aimed at navigating the space of possibilities by organizing and synthesizing these solution fragments. The nailer team focused on the energy storage, conversion, and delivery subproblems and had generated dozens of concept fragments for each subproblem. One approach to organizing and synthesizing these fragments would be to consider all of the possible combinations of the fragments associated with each subproblem; however, a little arithmetic reveals the impossibility of this approach. Given the three subproblems on which the team focused and an average of 15 fragments for each subproblem, the team would have to consider 3,375 combinations of fragments (15 x 15 x 15). This would be a daunting task for even the most enthusiastic team. Furthermore, the team would quickly discover that many of the combinations do not even make sense. Fortunately, there are two specific tools for managing this complexity and organizing the thinking of the team: the concept classification tree and the concept combination table. The classification tree helps the team divide the possible solutions into independent categories. The combination table guides the team in selectively considering combinations of fragments.
Concept Classification Tree
The concept classification tree is used to divide the entire space of possible solutions into several distinct classes that will facilitate comparison and pruning. An example of a tree for the nailer example is shown in Exhibit 7-7. The branches of this tree correspond to different energy sources.
The classification tree provides at least four important benefits:
1. Pruning of less promising branches: If by studying the classification tree the team is able to identify a solution approach that does not appear to have much merit, then this approach can be pruned and the team can focus its attention on the more promising branches of the tree. Pruning a branch of the tree requires some evaluation and judgment and should therefore be done carefully, but the reality of product development is that there are limited resources and that focusing the available resources on the most promising directions is an important success factor. For the nailer team, the nuclear energy source was pruned from consideration. Although the team had identified some very intriguing nuclear devices for use in powering artificial hearts, they felt that these devices would not be economically practical for at least a decade and would probably be hampered by regulatory requirements indefinitely.
2. Identification of independent approaches to the problem: Each branch of the tree can be considered a different approach to solving the overall problem. Some of these approaches may be almost completely independent of each other. In these cases, the team can cleanly divide its efforts among two or more individuals or task forces. When two approaches both look promising, this division of effort can reduce the complexity of the concept generation activities. It also may engender some healthy competition among the approaches under consideration. The nailer team found that both the chemical/explosive branch and the electrical branch appeared quite promising. They assigned these two ap proaches to two different subteams and pursued them independently for several weeks.
3. Exposure of inappropriate emphasis on certain branches: Once the tree is constructed, the team is able to reflect quickly on whether the effort applied to each branch has been appropriately allocated. The nailer team recognized that they had applied very little effort to thinking about hydraulic energy sources and conversion technologies. This recognition guided them to focus on this branch of the tree for a few days.
4. Refinement of the problem decomposition for a particular branch: Sometimes a problem decomposition can be usefully tailored to a particular approach to the problem. Consider the branch of the tree corresponding to the electrical energy source. Based on additional investigation of the nailing process, the team determined that the instantaneous power delivered during the nailing process was about 10,000 watts for a few milliseconds and so exceeds the power that is available from a wall outlet, a battery, or a fuel cell (of reasonable size, cost, and mass). They concluded, therefore, that energy must be accumulated over a substantial period of the nailing cycle (say 100 milliseconds) and then suddenly released to supply the required instantaneous power to drive the nail. This quick analysis led the team to add a subfunction ("accumulate translational energy") to their function diagram (see Exhibit 7-8). They chose to add the subfunction after the conversion of electrical energy to mechanical energy, but briefly considered the possibility of accumulating the energy in the electrical domain with a capacitor. This kind of
refinement of the function diagram is quite common as the team makes more assumptions about the approach and as more information is gathered.
The classification tree in Exhibit 7-7 shows the alternative solutions to the energy source subproblem. However, there are other possible trees. The team might have chosen to use a tree classifying the alternative solutions to the energy delivery subproblem , show ing branches for single impact, multiple impacts, or pushing. Trees can be constructed with branches corresponding to the solution fragments of any of the subproblems, but certain classifications are more useful. In general , a subproblem whose solution highly constrains the possible solutions to the remaining subproblems is a good candidate for a classification tree. For example, the choice of energy source (electrical , nuclear, pneumatic, etc.) constrains whether a motor or a piston-cylinder can be used to convert the energy to translational energy. In contrast, the choice of energy delivery mechanism (single impact,
multiple impacts, etc.) does not greatly constrain the solutions to the other subproblems. Reflection on which subproblem i likely to most highly constrain the solutions to the remaining subproblems will usually lead to one or two clear ways to construct the classification tree.