Applied Risk Analysis and Monte Carlo Methods

1. Applied Risk Analysis, Simulation and Software Development Ideally one would like to have a software tool for doing risk analysis, which combines influence diagrams, network models, discrete event models and system dynamic models. With such a hybrid t ool one could choose the model type that best suited the problem, and possibly integrate various types of models in a combined model. This is partly the idea behind the so-called Ptolomy-project developed at the CHESS Research Group (see http://chess.ee cs.berkeley.edu), University of California, Berkeley. I have been invited by one of the principal investigators at this center, Professor Edward A. Lee (see http://ptolemy.eecs.berkeley.edu/~eal/), to spend my sabbatical in his group. The purpose of this stay is to cooperate on developing this concept further. In particular, we will explore how the Ptolemy-platform can be used in relation to various problems in applied risk analysis, especially in connection to offshore related activities. 2. Monte Carl o Methods and Real Options One of the most active areas of research within the field of risk analysis today is real option analysis (ROA). A "real" option is an option "relating to things", i.e., fixed, permanent, or immovable objects, as opposed to fina ncial instruments like stocks and bonds. During my sabbatical, I plan to work on Monte Carlo-methods for solving optimisation problems related to real option analysis. There is currently a great potential for improvements in this area by employing more advanced techniques of simulation. Particularly, recent methods for analysing American options will be of relevance. While at the University of California, Berkeley, I will study this in further detail together with the CHESS research group. We will work on this both from a theoretical perspective and from an implementation point of view.