We present a model of network evolution in which the formation/deletion of links is coupled with the dynamics of the state variable of the nodes of the network. The model is originally intended to describe an economic network of agents engaged in knowledge production through knowledge exchange, although the framework can be generalized to other evolving networks. With respect to previous works, we improve by investigating how rich structures can emerge or collapse, depending on the individual rule of link formation/deletion used by selfish and boundedly rational agents. We characterize the emerging network topologies in terms of their efficiency and stability. The model reproduces qualitatively the main stylized facts of innovation networks, namely that they are sparse, locally dense and heterogeneous in degree.