Agriculture Reference
In-Depth Information
Some Optimization Exercises
conditions. Broiler house turn-round (7 days),
fixed production costs in Rand, R273/m
2
per
annum) and variable costs (R9.20/bird per
cycle) are illustrative. Mortality was set at 5%
to 35 days. Base revenue was generated at
R25/kg dressed weight.
Because circumstances differ to such an ex-
tent from producer to producer and from
country to country, examples relevant to one
set of circumstances are unlikely to be of
interest to others, so these examples of the
optimization process should not be regarded
as being universally applicable. However,
some illustrations of the use of the program
may be useful.
The first exercise has been mentioned
above. Fisher (2008) demonstrated that the
optimum balanced protein levels in the feed
differed depending on the way in which the
broilers were sold
(Table 13.2
)
. Using the
same dataset but different feed ingredient
prices (those applicable in 2006 and in
2008) he also demonstrated that these opti-
mum protein levels differed when the cost
of protein relative to energy changed, the
optimum often being higher than the recom-
mended level
(
Table 13.3
).
The exercise by Fisher (2008) was based
on actual performance, and the use of differ-
ent ingredient costs and revenues for the
various scenarios. Three additional exercis-
es are discussed below that are based on
simulated data using the EFG broiler growth
model and optimizer (EFG Software, 1995),
their purpose being to illustrate the effects of
changes to different objective functions or
variables on the optimum feed composition
and feeding schedule. Birds were reared to a
fixed weight of 2.3 kg or to 35 days and rev-
enue was generated on a processed basis only.
Feed ingredient availability and price were
set at some arbitrary but currently realistic
Optimizing amino acid contents in each
feed using different objective functions
In this exercise a three-stage feeding pro-
gramme was used, using 800 g starter/bird,
1200 g grower and offering a finisher until
the mean live weight of the mixed flock of
broilers reached 2.3 kg. All feeds had an en-
ergy content of
13
MJ ME/kg. Four objective
functions were chosen (scenarios A-D);
namely, maximizing margin over feed cost
(A), maximizing breast meat yield (B), min-
imizing feed conversion ratio (FCR) (C) and
minimizing nitrogen (N) excretion (kg N/flock
of
30,000
birds) (D).
Table 13.4
shows a sum-
mary of the optimum digestible lysine con-
tents in each of the three feeds offered, as
well as the predicted value of each of the
objectives at the optimum for each scenario.
The effect of choosing different object-
ive functions is to change the protein (shown
here as digestible lysine) contents in each of
the feeds on offer, with the result that the
profitability of the enterprise changes. The
objective of any broiler enterprise should
be to maximize profit, but in this example
costs and revenue were included in only
one of the objective functions (scenario A),
hence margin over food cost was lower by
comparison in the three other cases. In many
operations the aim appears to be to minimize
FCR, but when this is achieved (scenario C)
the profit margin may be seriously eroded,
as shown in
Table 13.4.
Clearly the differ-
ences in profitability depend on prevailing
circumstances, being larger or smaller than
indicated depending on the ratio between
ingredient costs and returns. In some coun-
tries the cost of removing N waste may be so
great that the lower performance and mar-
gin of broilers on scenario D, where N ex-
cretion is minimized, may be the most
acceptable and beneficial for the enterprise.
Table 13.3.
Level of balanced protein (%) relative to
Aviagen recommendations yielding the maximum
profit when birds were sold either at the farm gate or
processed, using feed ingredient prices applicable
in either 2006 or 2008. (From Fisher, 2008.)
Farm gate prices
Processed prices
Strain
2006
2008
2006
2008
Ross 708 M
100
115
130
130
Ross 708 F
90
115
115
115
Ross 308 M
100
115
115
130
Ross 308 F
90
115
100
115
