Return to ASU decision analysis resources page.EXPRESSION TREE is a MS-DOS program that allows rapid non-graphical construction and analysis of decision trees. It was originally developed in the late 1980s to aid decision analysis instruction, and it features ease of use, rapid computation speed, and minimal hardware requirements. When it was developed, there was no inexpensive full-featured decision tree or influence diagram program available for instructional use. While EXPRESSION TREE does not have the mouse-oriented graphical user interface of more recent programs, it still has at least one advantage over most student versions of commercial software: It does not have their limitations on model size or execution speed. It runs on IBM compatible computers operating under MS/PC-DOS Version 2 or later with any display and at least 256K memory. The program will also run in a DOS window under Windows.
Algebraic expressions can be used to determine the values of end points on a tree. These expressions can include numbers, arithmetic operators (+, -, /, log, etc.), logical operators (AND, OR, >, etc.), parameter names, and node names. If a parameter name is used (for example, TAX_RATE), then the value of the parameter is determined from a table which the user enters. If a node name is used, then it refers to a decision or chance node in the tree. The user enters different numerical values on the branches of that decision or chance node, and the appropriate value is automatically used for each end point on the tree. Exponential utility functions are also supported.
In addition, algebraic expressions including numbers, arithmetic and logical operators, and probability variable names can be used to determine the values of branch probabilities for any chance node. If a probability variable name is used (for example, P_HIGH_BID) then the value of this is determined from a table which the user enters.
The use of parameters, probability variables, and algebraic expressions often allows large decision trees to be rapidly entered and modified. It may only require entering a few distinct nodes to analyze a tree which has several dozen or even several hundred end points. In addition, automated sensitivity analysis can be conducted for any parameter, probability variable, or the risk tolerance for an exponential utility function.
Download Instructions: The EXPRESSION TREE package includes the program, a complete manual in Adobe PDF, and tutorial examples. This package is in the zip archive etree.zip (267,104 bytes). The manual ET.PDF included in this package can be displayed or printed using Adobe Acrobat Reader, which is available free from Adobe Systems Incorporated. Note: When you print from Acrobat Reader, uncheck the option “Fit to page” on the Print dialog to increase the size of the printed type.
Return to ASU decision analysis resources page.
Last updated August 1, 2001.