The
change of expert rules may lead to errors and conflicts. Different
from expert system, which is based on expert rules, case based reasoning
(CBR) [9] is dependent on typical scenarios. It is hard to
establish an excellent case base because of the small probability
of blackouts, and the difficulty of maintenance is also a bottleneck
for its application. Besides, mathematical programming [10] is
advantageous to obtain an optimal solution, but the application
is rare due to its sheer difficulty, huge solution space and long
execution time. With the development of computational intelligence,
some heuristic algorithms like genetic algorithm [11], artifi-
cial neutral network [12] and fuzzy theory [13] are introduced into
restoration field. However, the optimum of the methods cannot be
guaranteed, and these algorithms need to be improved in some
specific aspects respectively. Graph theory is also a popular algorithm
for restoration, the switch operation and network topology
is described with Petri net. This method is straightforward and
accessible in a small system, but a little too complex in large
systems [14]. The verification of constraints and disposal of uncertainties
need further improvement.
In addition, more algorithms like multi-agent [15] have been