Section: Application Domains

Agronomy

Quality control within agri-food chains, but also non-food chains relies on numerous criteria (environmental, economical, functional, sanitary quality, etc.). The objectives of quality are based on several actors. The current structure of chains is questioned as for system perenniality, protection of the environment, cost and energy. In all cases, the following questions have to be taken into account:

1. the actors' viewpoints are divergent, hence it is necessary to define reasoning mechanisms able to model and take into account the balance between viewpoints, and the risks and benefits they imply;

2. the successive steps involved in a chain, impacting the quality of end products, have limiting factors. Their improvement is a complex objective that has no simple solution;

3. data from literature are dispersed and scattered, which makes their use difficult.

These questions highlight the need for an integrated approach of agri-food chains, respectively with symbolic reasoning mechanisms, reverse engineering methods, and knowledge organization and modelling.

Our general objective is the conception of a decision support tool for the actors of an agri-food chain, in presence of contradictory viewpoints and priorities, including the concepts of gravity and certainty of a risk or a benefit. The first step is to build a knowledge-based system able to represent the different kinds of knowledge needed, and provided with consistency checking, querying and symbolic simulation mechanisms, which will allow to refine and validate the modelling.

Our results in 6.2 , 6.3 and 6.4 can be seen as theoretical requirements towards this objective.