Speciation as automatic categorical modularization
IEEE Transactions on Evolutionary Computation
| Hi-index | 0.00 |
In this paper, we propose a method to obtain strategy coalitions, whose confidences are adjusted by genetic algorithm to improve the generalization ability, in the process of co-evolutionary learning with a social game called Iterated Prisoner's Dilemma (IPD) game. Experimental results show that several better strategies can be obtained through strategy coalition, and evolutionary optimization of the confidence for strategies within coalition improves the generalization ability.