Optimising nutrient density
Given that the user wishes to retain a given ratio between the essential amino acids and energy, the program will optimise the nutrient density in each of the feeds in the feeding programme, by maximising the objective function over the entire growth period. As Fisher and Wilson (1974) have shown, the optimum nutrient density depends on such factors as sex, the ratio between input and output costs, and mixing and transport costs. These factors, and others, may be considered by the user in determining the optimum nutrient density of each of the feeds in the programme that will maximise overall performance or profitability.
To use this tool, select your optimiser flock, and go to the Nutrient density tab.
You will see an extra set of columns in the feeding schedule under an AMEn header. These are used to define the AMEn values that the optimiser should consider. There are two basic methods that can be used to arrive at a solution.
- Numerical method
- Grid method
If the grid method is used, the user enters the number of points to be considered for each feed. The interval size is calculated automatically using the Min/Max bounds given for each feed, and the number of points to be considered. In this way the user retains control over the accuracy obtained for each value in a solution.
Note that if there are many feeds, then a large number of points can result in a very slow solution, as each combination of values has to be solved. It is therefore better to start with a low number of points, and to use the information from such a solution to narrow the ranges and increase the accuracy for a subsequent solution.
The points that will be considered are listed on the Points tab.
This table shows the values that will be considered for each feed. The optimiser will consider all possible combinations of these points.
If the Refine soln. option is ticked, then the optimiser will do a final numerical solution using the best solution from the grid method as a starting point. This process is sometimes able to obtain a better solution than the grid method on its own.
If the numerical method is selected, then NAG (Numerical Algorithms Group) utilities will be used to solve the problem numerically. In this case, the columns that appear will be the Min and Max bounds (restricting the range of AMEn values that should be considered) as well as a Start AMEn column. This defines a starting point for the numerical method.
Note that this method is very sensitive to the starting point, and will not be able to find a solution if the starting point is not suitable.
See Comparison of the numerical and grid methods.
Once all inputs have been set up, click the Solve button on the bottom toolbar. Note that this button is only enabled if one input item of each type has been dragged down to the flock.