Detail

Probabilistic Constraint Reasoning in Continuous Domains

The present work proposes an extension to the classical continuous constraint approach to complement the interval bounded representation of uncertainty with a probabilistic characterization of the values distributions.
This extension provides an extra characterization of the uncertainty related with a given system by allowing to associate probability distributions with the variable domains. By reasoning with this new information it is possible to further characterize the solution scenarios with a likelihood value.
In the talk, a formalization of a Probabilistic Continuous Constraint Paradigm is presented and its main features are described, together with experimental results obtained with its application to inverse and reliability problems.