Sunday, August 29, 2010


Usefulness of Decision Tree Approach

The decision tree approach is extremely useful in handling the sequential investments. Working backwards from future to presents we are able to eliminate unprofitable branches and determine optimum decision at various decision points. The merits of the decision tree approach are:
  • Clarity. It clearly brings out the implicit assumptions and calculations
  • Graphic visualisation. It allows a decision maker to visualise assumptions and alternatives in graphic form, which is usually much easier to understand than the more abstract, analytical form.

However, the decision tree diagrams can become more and more complicated as the decision maker decides to include more alternatives and more variables and to look farther and farther in time. It is complicated even further if the analysis is extended to include interdependent alternatives and variables that are dependent upon one another; for example, sales volume depends on market share which depends on promotion expenses. The diagram itself quickly becomes cumbersome and calculations become very time consuming or almost impossible.

Sunday, August 1, 2010


Decision Trees for Sequential investment Decisions

In the simple accepted or reject decisions, we can use simple techniques. In practice, the present investment decisions may have implications for future investment decisions. Such complex investment decisions involve a sequence of decisions over time. It is argued that “since present choices modify future alternatives, industrial activity cannot be reduced to a single decision and must be viewed as a sequence of decisions extending from the present time into the future. If this notion of industrial activity as a sequence of decisions is accepted, we must view investment expenditures not as isolated period commitments, but as links in a chain of present and future commitments. An analytical technique to handle the sequential decisions is to employ decision trees. Here we shall illustrate the use of decision trees in analyzing and evaluating the sequential investments.

Steps in decision tree approach

A present decision depends upon future events, and the alternatives of a whole sequence of decisions in future are affected by the present decision as well as future events. Thus, the consequence of each decision is influenced by the outcome of a chance event. At the time of taking decisions, the outcome of the chance event is not known, but a probability distribution can be assigned to it. A decision tree is a graphic display of the relationship between a present decision and future events, future decisions and their consequences. The sequence of events is mapped out over time in a format similar to the branches of a tree.

While constructing and using a decision tree, some important steps should be considered:
  • Define investment. The investment proposal should be defined. Marketing, production or any other department may sponsor the proposal. It may be either to enter a new market or to produce a new product.
  • Identify decision alternatives. The decision alternatives should be clearly identified. For example if a company is thinking of building a plant to produce a new product, it may construct a large plant, a medium sized plant or a small plant initially and expand it later on or construct no plant. Each alternative will have different consequences’.
  • Draw a decision tree. The decision tree should be graphed indicating the decision points, chance even and other data. The relevant data such as the projected can flows, probability distribution, and the expected present value; should be located on the decision tree branches.
  • Analyze data. The result should be analyzed and the best alternative should be selected.