The Random part in network evolution

Economic behavior strives for efficiency. Therefore, also evolving network structures should be a result of such a goal-oriented behavior. Traditionally, networks were assumed to be only temporary phenomena, since the prevailing organizational forms that comply with the efficiency postulate are either firms or markets. Having a goal in mind, however, does not incur a set of unique choices of action, especially in situations under high uncertainty when engaging in invention networks. Consequently, there is no uniqueness in network structures. There is a random part in network evolution driven by generic mechanisms. A percolation model is used to model the generic development of invention networks. A Monte-Carlo simulation underlines the expectable patterns of network evolution. Moreover, it is tried to align the generic part of the story to the operant level where entrepreneurial behavior and market selection takes over the dominant role in network formation. -- R&D cooperation ; percolation theory ; knowledge diffusion ; networks.