Taking into account environmental criteria in animal feed formulation
Some databases can attribute environmental values to most raw materials used in animal feed production. These values, obtained through analysing the life cycle of the raw materials, can be used in formulation as they are additive variables. The impact of a mix of raw material will therefore be the sum of the impacts of each ingredient, weighed by the incorporation rate in the mix.
It is thus already possible to calculate, and limit, the environmental impact of the feed.
How are these environmental criteria taken into account?
In formulation software, this can be done by creating characteristics in raw materials that, as a standard nutrient, will describe the environmental values of the ingredient. For example, the carbon footprint linked to the production of the raw material, its impact of deforestation, and so on. Formulas will then be optimised on least cost and the environmental criteria will be evaluated, with eventually maximum constraints on some of the criteria.
Forcing limits to the formula will of course imply an overcost on the diet. The formulator will have to try, by tests and error, to find the best balance between the various criteria, while keeping an eye on the cost of the diet so as not let it increase excessively. The economic optimum is not necessarily the environmental optimum.
If environmental criteria have been aggregated in a single function, it can still be relevant to check the weight of each criterion individually.
The limits of current solutions
In current formulation software, it is not easy to consider multiple objectives in optimisation. The “economic function” of the optimisation is the price. It is the price that we want to minimise while respecting constraints of the formula. If we wanted to minimise the carbon footprint of the formula for example, it would mean going though a series of complicated process to replace the price of raw materials with their carbon footprint and setting the price as a nutrient in the raw material in order to put constraints on the price of the diet. For practical reasons, this could not be done as part of the daily work. Results would be fairly complicated to interpret.
A-Systems suggests new methods: the multi-objectives optimisation in Allix
Taking into account the environmental impact in single formulation
To allow considering several criteria in the optimisation of diets, A-Systems developed, in the formulation software Allix, the multi-objectives optimisation. This new optimisation allows formulators to minimise environmental impacts of feed formulation while still controlling the price of the diet. It is also possible to know what overcost can be accepted to allow minimising the environmental impact of the feed production. Based on factory and market context, the formulator can choose the weight to put on each criterion, as well as the cost of the feed.
Taking into account environmental impact in multi-formulation
A second methods is to minimise the environmental impact of the overall feed production of the factory. It is possible to calculate a multi-formulation of all formulas of the factory, each weighed by the tonnage planned to be produced. Environmental impact constraints will measure and limit the impact of all raw materials on pre-selected criteria. Furthermore, if the data was available, which is sadly not the case today, it would be possible to consider impacts linked to the production of the feed (grinding, granulation…).
Taking into account the environmental impact in the Feeding Plan
Through the rearing cycle, an animal will receive various types of feed, for example starter, grower, and finisher feeds. It is more relevant to work on the full life cycle of the animal than on individual feeds. This will allow the minimising not only of the impact of a single feed, but of the full rearing cycle, at least from a feeding perspective. That is why A-Systems now suggest the optimisation of “feeding plans” with multi-objectives formulation. It will be possible to minimise the economic costs of the diets, while finding the environmental optimum. Depending on the limitation posed by the agricultural sector, it is also possible to distribute the impact in a measured and optimal fashion through the rearing cycle.
These three new methods of optimisation developed by the A-Systems teams are innovative, and broaden considerably the scope of feed formulation. Indeed, we spoke here mainly of environmental impact. But the same tools can also answer other questions! For example, what would be the optimal level of selected nutritional criteria for an agreed upon overcost of the feed?