.
Economic Decision Trees 317
1: Build New
Product
8: Revenue
$100Klyear
7: Net Revenue
="$0
4: Net Revenue
Year!:;:: $100K
5: Net Revenlle
Year 1 "".$200K Year 2 ...n
6: Net Revenue
Year 1==$400K
9: Net Revenue
=.$600K/year
10: Net Revenue
==$400K/year
t== 0
FIGURE 10-3 Economic decision tree for new product.
t== (^1) t==2, ...,ex>
Figure 10-3illustrates that decision trees describethe problemby starting at the decision
that must be made and then adding chance and decisionnodesin the proper logical sequence.
Thus describing the problem starts at the first step and goes forward in time with sequences
of decision and chance nodes.
To make the decision, calculations begin with the final nodes in the tree. Since they
are the final nodes, enough infonnation is availableto evaluate them. At decision nodes the
criterion is either to maximize PW or to minimize EUAC. At chance nodes an expected
value for PW or EUAC is calculated.
Once all nodes that branch from a node have been evaluated, the originating node <;:an
be evaluated. If the originating node is a decision node, choose the branch with the best
PW or EUAC and place that value in the node. If the originating node is a chance node,
calculate the expected value and place that value in the node. This process "rolls back'"
values from the terminal nodes in the tree to the initial decision. Example 10-10 illustrates
this.
I
f
"
IiI
".!_
.J;
