Performance Evaluation of an Advanced Local Search
Evolutionary Algorithm
Anne Auger and Nikolaus Hansen
One natural question when testing performance of global
optimization algorithm is: how performances compare to a restart
local search algorithm. One purpose of this paper is to provide
results for such comparisons. To this end, the performances of a
restart (advanced) local-search strategy, the CMA-ES with small
initial step-size, are investigated on the 25 functions of the CEC
2005 real-parameter optimization test suit. The second aim is to
clarify the theoretical background of the performance criterion
proposed to quantitatively compare the search algorithms. The
theoretical analysis allows us to generalize the criterion
proposed and to define a new criterion that can be applied more
appropriate in a different context.
ERRATA:
1) End of Section 3, stopping criterion "noeffectcoord":
"Stop if adding 0.2-standard deviation in each coordinate does
change ..."
should be
"Stop if adding 0.2-standard deviation in any coordinate does
not change ..."
The following has been pointed out by Xiao-Min Hu from the Sun
Yat-sen University:
1) On function F4 the initial search point was
chosen by mistake as
Xup + (Xup-Xlow)*rand instead of
Xlow + (Xup-Xlow)*rand
for each coordinate. Xiao-Min Hu reported significantly
better results when using the corrected initialization.
2) On some function the boundary setting was omitted, but
this has most probably no impact on the results.