TY - GEN
T1 - On the adaptation of noise level for stochastic optimization
AU - Teytaud, O.
AU - Auger, A.
PY - 2007/12/1
Y1 - 2007/12/1
N2 - This paper deals with the optimization of noisy fitness functions, where the noise level can be reduced by increasing the computational effort. We theoretically investigate the question of the control of the noise level. We analyse two different schemes for an adaptive control and prove sufficient conditions ensuring the existence of an homogeneous Markov chain, which is the first step to prove linear convergence when dealing with non-noisy fitness functions. We experimentally validate the relevance of the homogeneity criterion. Large-scale experiments conclude to the efficiency in a difficult framework.
AB - This paper deals with the optimization of noisy fitness functions, where the noise level can be reduced by increasing the computational effort. We theoretically investigate the question of the control of the noise level. We analyse two different schemes for an adaptive control and prove sufficient conditions ensuring the existence of an homogeneous Markov chain, which is the first step to prove linear convergence when dealing with non-noisy fitness functions. We experimentally validate the relevance of the homogeneity criterion. Large-scale experiments conclude to the efficiency in a difficult framework.
UR - https://www.scopus.com/pages/publications/79955243436
U2 - 10.1109/CEC.2007.4424857
DO - 10.1109/CEC.2007.4424857
M3 - Conference contribution
AN - SCOPUS:79955243436
SN - 1424413400
SN - 9781424413409
T3 - 2007 IEEE Congress on Evolutionary Computation, CEC 2007
SP - 3027
EP - 3034
BT - 2007 IEEE Congress on Evolutionary Computation, CEC 2007
T2 - 2007 IEEE Congress on Evolutionary Computation, CEC 2007
Y2 - 25 September 2007 through 28 September 2007
ER -