Belief revision via Lamarckian evolution

We present a genetic algorithm for performing belief revision in a multi-agent environment. In this setting, different individuals are exposed to dierent experiences. This may happen because the world surrounding an agent changes over time or because we allow agents exploring dierent parts of the world. The algorithm permits the exchange of chromosomes from dierent agents and combines two dierent evolution strategies, one based on Darwin's and the other on Lamarck's evolutionary theory. Experiments on a problem of digital circuit diagnosis and on the n-queen problem show that the addition of the Lamarckian operator in the single agent case improves the tness of the best solution, even if in the digital circuit case the tness dierence is not signicant. Moreover, the experiments show that the distribution of constraints, even if it leads to a decrease of the tness of the best solution, does not produce a signicant dierence