Not much can be done to prevent the occurrence of a disaster, but one can be prepared in the
event of such a disaster. Pre-positioning of aid supplies has become a crucial part of disaster
management, providing the means to overcome the impact of a disaster and to reduce the
suffering and loss of life associated with these events.
Natural and man-made disasters are largely unpredictable, therefore disasters need to be
researched and their impact fully understood so that the aid supplies required to maintain
survival during and after disaster events will be available.
The member states within the South African Development Community (SADC) are the
countries of interest for this project, as research conducted on inventory pre-positioning for
disaster response in these countries has been insufficient. It is imperative to anticipate the needs
of disaster victims for potential disasters. This need is evaluated according to the types and
amounts of aid supplies required. All the natural and man-made disasters that have manifested
in these countries over the last 30 years are identified, as well as the appropriate aid supplies
required to successfully address the impact of such disasters.
General research contributions applied to humanitarian operations have progressed signif-
icantly, whilst operations research methods have not yet been comprehensively introduced to
cater for disaster-related problems.
A literature review examines the problems associated with the disaster preparedness phase
and the existing models used to solve these problems. This project proposes two generic mathe-
matical models that can be used to effectively determine the types and quantities of aid supplies
required in pre-positioning facilities within the SADC. A Disaster Recovery Plan (DRP) model
is formulated to enhance the survival capability of people in the region directly after the oc-
currence of any of the disasters identified and a stochastic inventory model is presented to
minimise the total cost of the aid supplies kept in inventory. Finally, various methods are
utilised to illustrate the functionality of the models.
Thesis (BEng. (Industrial and Systems Engineering))--University of Pretoria, 2010.