We present implemented concepts and algorithms for a simulation approach to decision evaluation with second-order belief distributions in a common framework for interval decision analysis. The rationale behind this work is that decision analysis with interval-valued probabilities and utilities may lead to overlapping expected utility intervals yielding difficulties in discriminating between alternatives. By allowing for second-order belief distributions over interval-valued utility and probability statements these difficulties may not only be remedied but will also allow for decision evaluation concepts and techniques providing additional insight into a decision problem. The approach is based upon sets of linear constraints together with generation of random probability distributions and utility values from implicitly stated uniform second-order belief distributions over the polytopes given from the constraints. The result is an interactive method for decision evaluation with second-order belief distributions, complementing earlier methods for decision evaluation with interval-valued probabilities and utilities. The method has been implemented for trial use in a user oriented decision analysis software. 1. Introduction During the later decades decision analysis with imprecise or incomplete information has received a lot of attention within the area of utility theory based decision analysis. Stemming from philosophical concerns regarding the ability of decision-making agents to provide precise estimates of probabilities and utilities, as well as pragmatic concerns regarding the applicability of decision analysis, several approaches have been suggested, for example, approaches based on sets of probability measures [1] and interval probabilities [2]. With respect to methods for practical decision evaluation with imprecise input statements, a number of methods have been developed and some of them have been also implemented in computer software tools. Early works include the approach to decision making with linear partial information about input statements [3]. This approach promotes the conservative -maximin decision rule together with the use of imprecise probabilities modelled by means of linear constraints, suggesting evaluation algorithms to obtain minimum expected values. However, imprecision is restricted to probability assignments, and it is not possible to allow for constraints between different alternatives. Related methods aim at investigating whether stochastic dominance holds between decision alternatives when probabilities (and weights) are
References
[1]
I. Levi, “On indeterminate probabilities,” Journal of Philosophy, vol. 71, no. 13, pp. 391–418, 1974.
[2]
K. Weichselberger, “The theory of interval-probability as a unifying concept for uncertainty,” in Proceedings of the 1st International Symposium on Imprecise Probabilities and Their Applications, 1999.
[3]
E. Kofler, Z. W. Kmietowicz, and A. D. Pearman, “Decision making with linear partial information (L.P.I.),” Journal of the Operational Research Society, vol. 35, no. 12, pp. 1079–1090, 1984.
[4]
A. D. Pearman, “Establishing dominance in multiattribute decision making using an ordered metric method,” Journal of the Operational Research Society, vol. 44, no. 5, pp. 461–469, 1993.
[5]
A. A. Salo and R. P. H？m？l？inen, “Preference ratios in multiattribute evaluation (PRIME)—elicitation and decision procedures under incomplete information,” IEEE Transactions on Systems, Man, and Cybernetics Part A:Systems and Humans., vol. 31, no. 6, pp. 533–545, 2001.
[6]
J. Mustajoki, R. P. H？m？l？inen, and A. Salo, “Decision support by interval SMART/SWING—incorporating imprecision in the SMART and SWING methods,” Decision Sciences, vol. 36, no. 2, pp. 317–339, 2005.
[7]
A. Jiménez, S. Ríos-Insua, and A. Mateos, “A generic multi-attribute analysis system,” Computers and Operations Research, vol. 33, no. 4, pp. 1081–1101, 2006.
[8]
P. G？rdenfors and N.-E. Sahlin, “Unreliable probabilities, risk taking, and decision making,” Synthese, vol. 53, no. 3, pp. 361–386, 1982.
[9]
L. Ekenberg and J. Thorbi？rnson, “Second-order decision analysis,” International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems, vol. 9, no. 1, pp. 13–37, 2001.
[10]
M. Danielson, L. Ekenberg, and A. Larsson, “Distribution of expected utility in decision trees,” International Journal of Approximate Reasoning, vol. 46, no. 2, pp. 387–407, 2007.
[11]
D. Sundgren, M. Danielson, and L. Ekenberg, “Warp effects on calculating interval probabilities,” International Journal of Approximate Reasoning, vol. 50, no. 9, pp. 1360–1368, 2009.
[12]
M. Danielson and L. Ekenberg, “A framework for analysing decisions under risk,” European Journal of Operational Research, vol. 104, no. 3, pp. 474–484, 1998.
[13]
M. Danielson, L. Ekenberg, J. Idefeldt, and A. Larsson, “Using a software tool for public decision analysis: the case of nacka municipality,” Decision Analysis, vol. 4, no. 2, pp. 76–90, 2007.
[14]
M. Danielson and L. Ekenberg, “Computing upper and lower bounds in interval decision trees,” European Journal of Operational Research, vol. 181, no. 2, pp. 808–816, 2007.
[15]
M. Danielson, “Generalized evaluation in decision analysis,” European Journal of Operational Research, vol. 162, no. 2, pp. 442–449, 2005.
[16]
T. Tervonen and R. Lahdelma, “Implementing stochastic multicriteria acceptability analysis,” European Journal of Operational Research, vol. 178, no. 2, pp. 500–513, 2007.
[17]
O. Caster and L. Ekenberg, “Combining second-order belief distributions with qualitative statements in decision analysis,” in Managing Safety of Heterogeneous Systems, Y. Ermoliev, M. Makowski, and K. Marti, Eds., vol. 658 of Lecture Notes in Economics and Mathematical Systems, pp. 67–87, 2012.
