Abstract: Problem comparing alternatives under interval uncertainty is studied. It is assumed that the compared alternatives have indicators of quality in the form of interval estimates. It is shown that mentioned problem cannot be unambiguously resolved by purely mathematical methods and requires using of decision maker’s preferences. From two possible situations of comparing, situation of unique choice and situation of repeated one, the first situation, which is typical for problems of forecasting, is analyzed. Quantitative measure of plausibility of implementation for tested hypothesis about preference of an alternative in comparison with others in their set and measure of risk connected with possible falsity of such hypothesis are introduced. It is shown that this risk is increased with increasing number of compared alternatives. Some methods to calculate risks as well as procedure of decision-making in the framework of a set of alternatives are proposed.

Keywords: comparing interval alternatives, dependence of the risk on number of alternatives, probability logic in comparing interval alternatives, methods and procedures of decision-making for the problem.

ACM Classification Keywords: H.1.2 Human information processing. G3 Distribution functions. I.2.3 Uncertainty, “fuzzy”, and probabilistic reasoning.

Link:

RISK BEHAVIOUR IN A SET OF INTERVAL ALTERNATIVES

Gennady Shepelev

http://www.foibg.com/ijima/vol04/ijima04-04-p01.pdf