Peter Durr, Claudio Mattiussi, Andrea Soltoggio, Dario Floreano (2008)
"Evolvability of Neuromodulated Learning for Robots."
Proceedings of ECSIS Symposium on Learning and Adaptive Behaviors for Robotic Systems, pp. 41--46
==================================================================================================


"When the agent's control system is realized by a neural network (NN), the learning can be implemented by modifying the synaptic weights w according to the following generalized Hebbian plasticity rule [7]."

     [7] Y. Niv, D. Joel, I. Meilijson, and E. Ruppin. Evolution of reinforcement learning in uncertain environments: A simple explanation for complex foraging behaviors. Adaptive Behavior, 10(1):5.24, Jan. 2002. => done



--------------------------------------------------------------------------------------------------
"From a robotics point of view the limitation of the results presented in previous experiments of neuromodulatory evolution such as [7, 12, 11] is the use of simplified tasks based on grid-like worlds and a choice between a finite, small set of actions."

"Thus, as in the grid world experiments presented in [11], the artificial neural network must adapt to the changing position of the higher reward and change its strategy in order to gain maximal fitness."


     [11] A. Soltoggio, J. A. Bullinaria, C. Mattiussi, P. Durr, and D. Floreano. 
Evolutionary Advantages of Neuromodulated Plasticity in Dynamic, Reward-based Scenarios.
In Artificial Life XI: Proceedings of the Eleventh International Conference on the Simulation and Synthesis of Living Systems, 2008. =>  done

     [12] A. Soltoggio, P. Durr, C. Mattiussi, and D. Floreano. Evolving neuromodulatory topologies for reinforcement learninglike problems. In Proceedings of the 2007 Congress on Evolutionary Computation. IEEE Press, 2007. => done




--------------------------------------------------------------------------------------------------
"For more details on the algorithm see [5]."

     [5] C. Mattiussi and D. Floreano. Analog genetic encoding for the evolution of circuits and networks. IEEE Transaction on Evolutionary Computation, 11(5):596.607, Oct. 2007. => done