LAUSR

We investigate the dependence of knowledge spread from the position of experts (highly knowledgeable agents) within networks. Moreover, we examine whether the influence of experts is conditioned by the way agents select other agents for knowledge acquisition. In this perspective, we compare the emerging knowledge spread for 5 representative policies for positioning experts, namely: random (at nodes selected with equal probability) and central (at nodes with high centrality: degree, closeness, betweenness, eigen-centrality). Each agent selects an in-neighbor for knowledge acquisition, excluding (filtering out) agents of lower knowledge. As “selections” may be implemented before or after “filtering”, we also compare the conventional wisdom of “Selection” before “Filtering” with the reverse order of “Selection” after “Filtering”. We perform simulations for the 5 policies of positioning experts in 3 representative classes of networks, namely: random, small-world, and scale-free. The key finding is that any policy for positioning experts (even random) in networks where agents “Select” after “Filtering”, results in significantly faster knowledge attainment (up to 70%), compared to any policy for positioning experts when agents “Select” before “Filtering”. Therefore, “Selecting” after “Filtering” is much more effective Knowledge Management Strategy, compared to just placing experts in key positions, without caring how the agents select other agents for knowledge acquisition.