CONTINUOUS TREES AND NEVADA SIMULATION - A QUADRATURE APPROACH TO MODELING CONTINUOUS RANDOM-VARIABLES IN DECISION-ANALYSIS

This paper introduces an improved technique for modeling risk and deci sion problems that have continuous random variables and probabilistic dependence. Variables are modeled with mixtures of four-parameter rand om variables, called ''continuous trees.'' Functions of random Variabl es are calculated using gaussian quadrature in a manner called ''NEVAD A simulation'' (NumErical integration of Variance And probabilistic De pendence Analyzer). This technique is compared with traditional decisi on-tree modeling in terms of analytic technique, solution-time complex ity, and accuracy. NEVADA simulation takes advantage of the probabilis tic independence in a decision problem while allowing for probabilisti c dependence to achieve polynomial computational-time complexity for m any decision problems. it improves on the accuracy of traditional deci sion trees by employing larger approximations than traditional decisio n analysis. It improves on traditional decision analysis by modeling c ontinuous variables with continuous, rather than discrete, distributio ns, A Bayesian analysis using a mixed discrete-continuous probability distribution for cigarette smoking rate is presented.