Back-Propagation Neural Network Architecture for Solving the Double Dummy Bridge Problem in Contract Bridge

Contract Bridge is an intelligent game, which increases the expose with multiple skills and knowledge because no player knows exactly what moves other players are capable of making. The ´Bridge´, being a game of imperfect information, is to be equally well defined, since the outcome at any intermediate stage is purely based on the decision made on the immediate preceding stage. The credits accumulated by one pair of bridge players towards the target in a fixed number of ´tricks´ is called Double Dummy Bridge Problem. The Back-propagation neural network architecture is used to take the best tricks in Double Dummy Bridge Problem. In summary, the study described in this paper provides a detailed comparison between two different activation functions which were used to train and test the data, hence their behavior in different situations.