Automated verification of disaster plans in incident management

Purpose – The purpose of this paper is to create a formal specification language for disaster plans in order to remove possible inconsistencies between disaster plans, and to enable the automated verification of properties from such plans against logs of actual incidents. Design/methodology/approach – Different types of properties in disaster plans have been identified and formalized using order-sorted predicate logic, enabling automated comparison of plans and verification of such properties against logs by means of software tools. Actual disaster plans and logs have been used as a case study to show the working of the approach. Findings – The automated approach can be used quite easily and result in important findings. For the case study disaster plans crucial differences were found that could have catastrophic consequences. Furthermore, it is shown that in the logs of a well-known incident the disaster plan was not followed. Practical implications – If the approach is introduced in practice, disaster plans would be stored in a formal format, enabling the automated comparison of disaster plans, and immediate detection of derivation from a disaster plan in case of an incident. Originality/value – Other literature about the formal modelling of disaster plans that includes both structural and dynamical aspects and allows representation of organizational structure at multiple aggregation levels has not been found. Nor has comparing the disaster plans using such a formal model, and using the model of the disaster plan to check empirical traces for compliance with this plan, been addressed in prior literature.