Risk-based decision making in complex systems: The ALBA method

Abstract

Making decisions in complex systems it is a complicated task to accomplish. As complexity and uncertainty increase, the use of scenarios to exploring that uncertainty becomes essential to support decision makers. The difficulty associated with the combinatorial need imposed by complex systems requires methods and tools to unburden analysts from the cumbersome task of manually deriving scenarios and the awkward one of properly managing them. The paper presents how the Artificial Logic Bayesian Algorithm (ALBA) method, thanks to the use of artificial logic (or, more formally, the Logic-based Artificial Intelligence), allows for analysts both to build complete partitions (i.e., complete sets of mutually exclusive choices) by “only” defining the logical and stochastic correlations amongst the selected elective random variables (leaving to the algorithm the burden to create the complete partition), and to nimbly managing scenarios.