Whenever people try to solve a problem, they resort to approximate data because of the incomplete knowledge they have of their surroundings. Their reasoning is based rather on qualitative standards than on quantitative ones. Human beings think in a fuzzy way. Instead of saying “this car is driving at 84 mph”, people usually say “This car is driving fast or very fast”. They think according to a “fuzzy logic”. Fuzzy logic – which was defined according to the theory of fuzzy sets introduced by L.A. Zadeh – is aimed at turning qualitative values into quantitative ones through a strict conceptual framework. After analyzing all the fuzzy data which make up for its reflection, our brain actually makes a decision which becomes totally real, accurate and involving.
Many applications of fuzzy logic have been developed on household appliances or on consumer electronic goods, where the lack of accurate data makes automation by usual methods impossible. It is in Japan that fuzzy logic was the most successful. From 1980 onwards, washing machines without settings or autofocus photo cameras using fuzzy approach actually appeared. Having proved reliable in other fields such as finance or medical diagnosis, fuzzy logic has become today a reality that can be applied to human behaviors so that they can be modeled at best while taking their uncertainties into account.
When it comes to life insurance, the role of fuzzy logic is undoubtedly useful. This technique allows to refine and to generalize the ACPR curves of short-term takeovers within the framework of Best Estimate patterns for Euro contracts. In addition, fuzzy logic enables to express more clearly the rate policy on Euro contracts than algorithms unintelligible to company managers.
It is this context that the following master thesis (written in French) has been conducted :