It is crucial to appropriately allocate emergency power sources for reducing the outage cost to the greatest extent in emergency circumstances. Existing publications on emergency power sources are focused on optimal scheduling of emergency power sources in urban areas

while less attention has been paid to reasonable allocations of emergency power sources and the impacts of uncertain load demands and dispatching time periods of emergency power sources caused by some factors such as traffic jam and road damage have not yet been addressed. Given this background

an optimization model for allocating emergency power sources is presented with uncertain load demands and dispatching time periods considered

with an objective of maximizing the social welfare. The well-established radial interval represented by the fluctuation radius for the boundaries of an interval value is employed to more appropriately characterize the uncertainties involved. The optimization model is next transformed into a pessimistic sub-model and an optimistic sub-model for speeding the solving process

and each of the sub-models is then solved by a robust optimization method. Finally

case studies are served for demonstrating the developed model and method for allocating emergency power sources.