A Lamarckian Approach for Neural Network Training
Neural Processing Letters
Belief Revision by Lamarckian Evolution
Proceedings of the EvoWorkshops on Applications of Evolutionary Computing
On the adaptive disadvantage of Lamarckianism in rapidly changing environments
ECAL'07 Proceedings of the 9th European conference on Advances in artificial life
| Hi-index | 0.00 |
It is widely recognised that many species adapt to complex and dynamic environments, but it is no longer accepted that an organism passes characteristics acquired during its lifetime to its offspring. However, in evolutionary computation such Lamarckian inheritance can be useful. Simulations of the benefits of Lamarckian inheritance have been reported in the literature. However, in this paper we present the first formal proof that Lamarckian inheritance can dominate more traditional individual (non-inheritable) learning. We present a parameterised model that can demonstrate conditions in which different inheritance types perform best, which we empirically validate.