1 DETERMINING CRITERIA WEIGHTS AS A FUNCTION OF THEIR RANKS IN MULTIPLE-CRITERIA DECISION MAKING 1 Dr. Hesham K. Alfares Dr. Salih O. Duffuaa Systems Engineering Department King Fahd University of Petroleum & Minerals Dhahran, Saudi Arabia ABSTRACT The problem of assigning weights from ordinal ranks appears in many contexts in multi-criteria decision making. In this paper, we present an empirical methodology for converting an ordinal ranking of a number of criteria into numerical weights. Based on this methodology, the weight for each criterion is expressed as a simple mathematical function of its rank and the total number of criteria. The proposed methodology is empirically developed, evaluated, and validated based on a set of experiments involving university students and faculty members. The proposed method is compared with well-known methods in the literature and has shown superiority in assigning criteria weights from ordinal ranks. INTRODUCTION In multiple criteria decision making, several methods are used to determine the relative criteria weights. These methods depend on the input of the decision maker(s), i.e., the approach used to compare the different criteria. In many situations, the only input provided by each decision maker is a list (in the order of priority) the factors they consider most relevant for evaluating and comparing the applicable alternatives. As an example the factors used to assign an overall audit score for a quality or maintenance system are ranked by the auditing team and a weight for each factor must be obtained from the given ranks. In goal programming, if the different objectives are ranked, their relative weights can be determined and used to combine multiple objectives into a single objective. This paper presents an empirically developed methodology to convert criteria ranks into relative criteria weights, using real-life experiments that involve surveys of university students and faculty members. In these experiments., participants were first asked to list the relevant factors in the order of importance, and then asked to give a weight for each factor based on its importance in their point of view. In other words they had to provide a weight for each factor that matches the given rank (the highest weight must be given to the first-ranked factor). Using regression and statistical analysis, a methodology. That best fits the experimental data and minimizes the errors is recommended for general use in assigning weights. In order to validate the proposed methodology, a second set of experiments involving another sample of students and a different set of criteria was subsequently conducted. 1 Acknowledgments: The authors would like to acknowledge the support provided by King Fahd University of Petroleum & Minerals for conducting this research. Appreciation is also due to Mr. Saleh Al-Duwais for help in data collection and analysis.