The LIMID software system from Esthauge Decision Knowledge is a full Java software tool for construction and manipulation of graphical models in the form of Bayesian networks and decision graphs. It consists of an application programming interface for embedding with other Java software tools and a graphical user interface for visual construction, editing and maintenance of models.

For a general introduction to Bayesian networks and decision graphs, reference is made to Jensen & Nielsen (2007). Further details may be found in Cowell et al. (2007).

All implemented algorithms are based on construction of so-called junction trees after moralization and triangulation of the model specified by the user. The algorithm used for optimization in decision graphs is single-policy updating as described by Lauritzen and Nilsson (2001).

Anybody who needs a software tool for discrete Bayesian networks or decision graphs. The Java API makes it particularly well suited for embedding in multi-platform applications.

The facilities for visual display of moral graphs, triangulated graphs and junction trees make it the ideal tool for teaching.