Abstract
In this paper we first assume a peer-to-peer (P2P) network composing nodes with a variety of functions and then propose a method for each node to create new functions by combining others' functions in a P2P network. In the proposed method, a node (user) that desires a new function first creates a candidate of new function combining others' functions by using a source routing method that determines a next hop node based on numerical operations. Next, the node evaluates the created new function candidate and then genetic operators of a genetic algorithm adjusts parameter values used by the source routing method based on the user's evaluation value. Thus, new functions desired by a node are created by repeating the source routing for creating a candidate of new function, evaluation of the new function candidate by a node, and adjustment of the routing parameters by genetic operators. We show basic effectiveness of the proposed method through simulations.
