Supplier Selection Under Conditions of Uncertainty

2. Research Objective

The purpose of this research is to investigate the process that is commonly used for calculation of a “score” for each supplier using a known number of attributes. The score is then used for selection of suppliers. Many investigators have used the following linear model to calculate this score.

where S_j is the score for the j^th supplier

W_i is the weight or relative importance of i^th attribute

X_ij is the value of i^th attribute for j^th supplier

n is the number of attributes

Use of equation 1 requires that the values of both W_i and X_ij are known. Attention is mainly focused on the process by which the values of X_ij may be obtained. It is worth mentioning that it is very difficult to estimate these values precisely as these are routinely measured on an ordinal scale.

Classical decision-making problems deal with the selection of “best” alternative from a list of alternatives. Selection of a supplier is clearly a type of decision-making problem. In decision making theory, the selection of “best” alternative can be done under three different conditions: certainty, risk, and uncertainty. Certainty assumes that an outcome can take only one value. Underlying assumption under conditions of risk is that probability function of outcomes follows a known probability function. As an example, if probability function is “normal” then its mean and variance are known or can be estimated. In uncertainty, the decision maker knows the outcomes, but probability of each outcome is not known.