J. Yen et al., A HYBRID APPROACH TO MODELING METABOLIC SYSTEMS USING A GENETIC ALGORITHM AND SIMPLEX-METHOD, IEEE transactions on systems, man and cybernetics. Part B. Cybernetics, 28(2), 1998, pp. 173-191
One of the main obstacles in applying genetic algorithms (GA's) to com
plex problems has been the high computational cost due to their slow c
onvergence rate, We encountered such a difficulty in our attempt to us
e the classical GA for estimating parameters of a metabolic model, To
alleviate this difficulty, we developed a hybrid approach that combine
s a GA with a stochastic variant of the simplex method in function opt
imization, Our motivation for developing the stochastic simplex method
is to introduce a cost-effective exploration component into the conve
ntional simplex method. In an attempt to make effective use of the sim
plex operation in a hybrid GA framework, we used an elite-based hybrid
architecture that applies one simplex step to a top portion of the ra
nked population, We compared our approach with five alternative optimi
zation techniques including a simplex-GA hybrid independently develope
d by Renders-Bersini (R-B) and adaptive simulated annealing (ASA), Our
empirical evaluations showed that our hybrid approach for the metabol
ic modeling problem outperformed all other techniques in terms of accu
racy and convergence rate, We used two additional function optimizatio
n problems to compare our approach with the five alternative methods,
For a sin function maximization problem, our hybrid approach yields th
e fastest convergence rate without sacrificing the accuracy of the sol
ution found. For De Jong's F5 function minimization problem, our hybri
d approach is the second best (next to ASA), Overall, these tests show
ed that our hybrid approach is an effective and robust optimization te
chnique. We further conducted an empirical study to identify major fac
tors that affect the performance of the hybrid approach, The study ind
icated that 1) our elite-based hybrid GA architecture contributes sign
ificantly to the performance improvement and 2) the probabilistic simp
lex is more cost-effective for our hybrid architecture than is the con
ventional simplex, By analyzing the performance of the hybrid approach
for the metabolic modeling problem, we hypothesized that the hybrid a
pproach is particularly suitable for solving complex optimization prob
lems the variables of which vary widely in their sensitivity to the ob
jective function.