Introduction to Utility Evaluator

Optimization is an important technique in engineering design, but traditional optimization is not suitable to handle a multiobjective problem directly, Utility theory provides an analytical way to aid our decisions in engineering design. By exchanging from objective to attribute and expressing these attitudes mathematically, a utility-based attribute function (Utility Function) can be set up to describe the attitude of a decision maker with regard to his/her preference for consequences and lotteries to a certain attribute to facility the decision maker find the best design alternative.

In engineering design, Strict monotonicity is a very reasonable characteristic for a utility function. In short, Two possible conditions exist: Monotonic Increase (the more the better) and Monotonic Decrease (the less the better) which reflect utility change trend with the attribute increase. There are three basic risk attitudes toward one attribute: Risk Averse, Risk Prone and Risk  Neutral which illustrate the degree of the decision maker's confidence in being successful. For convenience, we commonly set the best preferred attribute level with utility "1" and least with "0". In practice, the utility with an exponential shape often meets most of the conditions and even the conditions are not met, the exponential formulation can provide an adequate approximation in a decision analysis.

This Utility Evaluator can be used for single-attribute utility analysis or determine certain parameters of a multiattribute utility function. It can calculate risk attitude constants and scaling constant, the utility for a specified attribute level or the attribute level for a specified utility. Also, a continuous graph is shown to clearly illustrate decision maker's risk attitude. Now we are trying to explore other utility forms to handle non-monotonic situations and target problems.

In this program, the first three lines are the basic information for the decision maker to a certain attribute. The next three lines are formed as a lottery question. It needs you to provide Certainty Equivalent for a lottery which you are assumed to know the best and least preferred attribute level. After finish these basic data, click "Evaluate" button, the program will show a graph in right-hand corner with an exponential coefficient in the message box. Then you can continue to choose the checkbox depending on if you need to find attribute level with a given utility or find utility with a given attribute level. The message box and output box will show error information if there are mistakes in data input. If you are not familiar with Utility Assessment procedures or Lottery Question, you can double click corresponding area, then relevant information is available immediately.

NOTE:  There are totally three textfields which are not editable, 1-Probability; Message Box and Output. They show computation results.
We use Normalized Exponential Utility Formulation in this program.

Now Let's go!

Utility Evaluator               Utility Theory Introduction                   Back to Lab