We want to provide advice to farmers, food and feed industry and decision makers in assessing the co-existence of GM and non-GM supply chains in a given environment and process. For example, in the production of starch from maize, there are many factors to be considered: the kernels used, transport, quantity of the GMO content, harvest cleaning and processing, starch refining and more. Some of these factors are highly complex. The software, which we call a decision support system (DSS), will consider all those factors and will give certain recommendations to the food and feed producers.
The aim is to develop software for modelling and simulating food and feed production chains. Within the SIGMEA project financed by the 6th EU Framework Programme, we already developed a decision support system called SMAC Advisor for the assessment of maize co-existence at the field level (see background information). Now we believe that we can use a similar methodological approach and similar computer tools to develop the Co-Extra DSS.
The DSS provides support for your decision-making. It won’t make decisions on your behalf, but would offer information with which you can decide better. At the basic level, the system will assess the GMO content in your product and associated production costs in some specific situations. Furthermore, the system will allow you to change some characteristics of your production parameters to see the potential effects. At a global level, the system will give a view on the whole production chains; stakeholders will be able to assess the effects of other people’s decisions on their processes at earlier production stages, and will see the effects of their decisions on other processes. Additionally, there is a tremendous explanatory and educational value of such a DSS - for instance, for the training of managers and the decision-making of authorities responsible for the food and feed safety.
Due to a matter of time, we will initially concentrate our efforts on the maize processing chain. From our experience, it takes about two to three months to develop a single reasonably complex decision model. There is also time needed for software development, integration, testing, etc. However, we will build the system to be as flexible as possible. This means that it will be able to represent and simulate many different processing chains with their specific procedures and operational steps.
Indirectly, at first. Stakeholders will become aware of the existence and availability of such a useful tool and will be able to assess its functionality. They can see how it might fit for their own purposes and, probably, will want to initiate the development of the missing components for their chain. In this case, I am sure they also can benefit from past experience and can reuse components which already are developed.
I am expecting the most difficulties in elaborating the decision models, which are developed in collaboration with experts in the processing chains. The decision problem is discussed and knowledge is collected, formalised and eventually represented in form of parameters, their structure, decision rules, probabilities, various formulae, etc. This process is always difficult and unpredictable, and it does take time. Face-to-face meetings are still indispensable in spite of all modern communication technologies. Thus, I am expecting to spend quite some time meeting with Co-Extra leading experts and discussing the ways to incorporate their knowledge into the DSS models.
Eventually, yes. I am quite sure we won’t be able to capture national variations right from the beginning, but we will gradually introduce them into the system. The key requirement is that the system and its components be sufficiently flexible to facilitate gradual improvement, and we achieve that by using our modular approach. In order to describe national variations, we might introduce special model parameters that differentiate between different national systems. In one of the SIGMEA models, for example, we already used this approach to deal with national variations of “regulatory environments”. However, when national differences are too large to be treated in this way, we might need to develop different models. For example, if it turns out that English starch production considerably differs from the French one.
We haven’t discussed this yet. Personally, I am in favour of bringing in stakeholders as early as possible. Usually in DSS development, this improves the quality of the final product and improves the chances for its adoption. On the other hand, this might complicate matters and mislead the development into focusing on too-specific problems of involved stakeholders. So, we still have to decide how to handle this issue. During the project, we will definitely seek stakeholders’ assessments of the DSS.