Risk and Duality in Multidimensions Bartlomiej Blaszczyszyn Mathematical Institute, Univers.pdfVIP

Risk and Duality in MultidimensionsBart lomiej B laszczyszynMathematical Institute,University of Wroc lawpl. Grunwaldzki 2/4,50-384 Wroc law, Poland;e-mail: blaszcz@math.uni.wroc.pl Karl SigmanyDepartment of Industrial Engineering andOperations Research, Columbia University,500 West 120th Street, MC:4704,NY, NY 10027, USA;email: sigman@ieor.columbia.eduMarch 31, 1999AbstractWe present, in discrete time, general-state-space dualities between content andinsurance risk processes that generalize the stationary recursive duality of Asmussenand Sigman [3] and the Markovian duality of Siegmund [12] (both of which areone-dimensional). The main idea is to allow a risk process to be set-valued, andto dene ruin as the rst time that the risk process becomes the whole space. Therisk process can also become innitely rich which means that it eventually takeson the empty set as its value. In the Markovian case, we utilize stochastic ge-ometry tools to construct a Markov transition kernel on the space of closed sets.Our results connect with strong stationary duality of Diaconis and Fill [4]. Asa motivating example, in multidimensional Euclidean space our approach yields adual risk process for Kiefer-Wolfowitz workload in the classic G=G=c queue, and weinclude a simulation study of this dual to obtain estimates for the ruin probabilities.Keywords: Choquet capacity, content process, Markov process, random closedset, risk process, set-valued process, space law, stochastic geometry, stochastic re-cursion, strong stationary dual.AMS 1991 Subject Classication; Primary: 60J05, 60K25, Secondary: 90A461 IntroductionIn one-dimensional Euclidean space a nice duality theory has been developed between\content like processes (such as storage or queues) and insurance \risk like processes.The major part of this research was done while the author was visiting the IEOR Department atColumbia University, and funded by the Kosciuszko Foundation and the Center for Applied Probability(CAP) at Colum