Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

BDD complexity reduction by component fault trees

 
: Kaiser, B.; Zocher, A.

Kolowrocki, K.:
Advances in Safety and Reliability. Proceedings of the European Safety and Reliability Conference, ESREL 2005. Vol.1
Leiden: Balkema Publishers, 2005
ISBN: 0-415-38340-4
S.1011-1019 : Ill., Lit.
European Safety and Reliability Conference (ESREL) <16, 2005, Gdynia; Gdansk; Sopot>
Englisch
Konferenzbeitrag
Fraunhofer IESE ()
safety analysis; reliability analysis; fault tree analysis; component fault tree; ESSaRel

Abstract
The analysis of Fault Trees (FTs) is usually performed by transformation into Binary Decision Diagrams (BDDs). The size of a BDD depends heavily on the order of its variables. Different approaches exist to optimise the variable order by reordering or applying heuristics. We present a new approach that is based on Component Fault Trees (CFTs), an FT extension that allows partitioning FTs according to the component architecture of a system, even if this does not result in independent sub-trees. CFTs are connected by input and output ports and can be decomposed hierarchically. CFTs turned out to be suitable frames for variable order optimisations for several reasons: First, components usually contain only a few variables so that order optimisation is practicable. Second, variables within a component are usually stronger coupled to each other than to variables outside, which impacts BDD complexity. Third, as safety or reliability critical systems often use redundant instances of the same component type, it is possible to perform the optimisation once and take profit of it several times. We explain the ordering algorithm, compare different variants by some case studies and give examples for the savings in effort.

: http://publica.fraunhofer.de/dokumente/N-34778.html