Abstract:

Investment decisions are very important problems for the companies to survive in the global world. Decision problems are hard because of including complexity, inherent uncertainty, multiple objectives and different perspectives lead to different conclusions. Decision trees are excellent tools in order to help making decisions. Decision makers can easily see alternative decisions and the implications of taking those decisions with constructing decision trees.

In this study, an investment problem associated with purchasing machine is solved by using decision trees method. The problem includes some kind of uncertainties; randomness, fuzziness and random fuzziness. There are three different stage associated with the demand; high, medium and low. The probabilities of the stages are calculated from the past data however management of the firm’s thinks that each of the probabilities has different possibilities considered the political developments, competitor’s strategy and etc. Therefore demand stage is modeled as a discrete random fuzzy variable. Similar to the demand stage variable, the demand quantity is including randomness and fuzziness. Demand quantity follows the normal distribution as a result of the analyzing past sales however the mean parameter of the normal distribution can be managed with performing different marketing strategies. Therefore demand quantity is modeled as a random fuzzy variable (normal distribution with fuzzy parameter).

This study shows that decision makers have an opportunity to manage the probability and these must be incorporated to the problem structure. In this manner decision makers can behave flexible and take more realistic decisions.

Keywords:

investment problem, decision trees, random fuzzy variable.