Analysis of pollution monitoring by expert systems

Abstract

Pollution monitoring can provide an important aid in the choice of the strategy to control the level of some dangerous elements, whether in water or in the air. The difficulties of detecting polluting sources from experimental data are related not only to the adoption of systematic and suitable measuring procedure, but also to a correct management of the available information. From the theoretical point of view, the use of simplified models, coupled with classical regularization techniques, shows that, in general, the problem is badly posed and consequently, numerically ill-conditioned. Hence the possibility of using expert systems algorithms, introducing further qualitative information, improves the reliability of the solutions. In particular, this paper deals with the utilization of fuzzy optimization algorithms: fuzzy theory supplies a formal reasoning technique, which proposes solutions that are real consequences of the premises.

An actual example of such a method is described, making reference to the computation of the distribution of polluting sources from ground concentration data. The inverse problem is first solved using traditional procedures, showing that the distributed sources are not recognized. Afterwards, different results obtained from various algorithms derived from the assumed a-priori knowledge are examined. In this case, it is possible to obtain a more realistic situation of the pollution sources, inside the boundaries of the controlled area.