Rps. Hart et al., OPTIMIZATION TECHNIQUES FOR A COMPUTER-SIMULATION OF A PASTORAL DAIRYFARM, Computers and electronics in agriculture, 19(2), 1998, pp. 129-153
This paper compares different methods of optimising the management var
iables in UDDER, a commercially-available computer simulation model of
a pastoral dairy farm. The emphasis is on identifying the best optimi
sation strategy for this complex multi-dimensional system, taking the
simulation model as a given constant. The optimisation methods studied
are based on significantly different principles, with differing stren
gths and weaknesses: two hill-climbing algorithms (Nelder-Mead simplex
and Powell's direction set), and a genetic algorithm (GA). Rather tha
n examine all facets of dairy farm management, a single problem is opt
imised-that of maximising milkfat production while maintaining the hea
lth of the herd and pasture. The results show that while the GA can de
termine good regions within the search space quickly, it is considerab
ly slower than either hill-climber at finding the optimal point within
that region. The hillclimbers, in contrast, are fast but have a tende
ncy to get trapped on local maxima and thus fail to find the true opti
mum. This led to the development of a hybrid algorithm which utilises
the initial global search of the GA, followed by the more efficient lo
cal search of a hill-climber. This hybrid algorithm discovered near-op
timal points much more quickly than the GA, and with more reliability
than the hill-climber. (C) 1998 Elsevier Science B.V.