A scenario-driven decision support system for serious crime investigation

Authors Organisations
Type Article
Original languageEnglish
Pages (from-to)87-117
Number of pages31
JournalLaw, Probability and Risk
Issue number2
Publication statusPublished - 16 Jan 2007
Permanent link
Show download statistics
View graph of relations
Citation formats


Consideration of a wide range of plausible crime scenarios during any crime investigation is important to seek convincing evidence and hence to minimize the likelihood of miscarriages of justice. It is equally important for crime investigators to be able to employ effective and efficient evidence-collection strategies that are likely to produce the most conclusive information under limited available resources. An intelligent decision support system that can assist human investigators by automatically constructing plausible scenarios, and reasoning with the likely best investigating actions will clearly be very helpful in addressing these challenging problems. This paper presents a system for creating scenario spaces from given evidence, based on an integrated application of techniques for compositional modelling and Bayesian network-based evidence evaluation. Methods of analysis are also provided by the use of entropy to exploit the synthesized scenario spaces in order to prioritize investigating actions and hypotheses. These theoretical developments are illustrated by realistic examples of serious crime investigation.


  • crime investigation, decision support, scenario generation, scenario fragments, Bayesian networks, evidence evaluation, conditional independence, system architecture, entropy