Sean Gourley: The Costs and Benefits of Network Approaches
Intelligent Agents Adaptive Systems – centralised networks and the effects of congestion.
Examples: Small World network – six degrees of separation and ordered network with a few random links. Scale Free network characterised by most nodes having very few connections but a few having many – like the www with hubs and super-hubs. Exponential network – roads.
The Hub & Spoke Network – when is it best to go around the outside and when to go through the hub. Slime Mould and Mycelial fungus do this all the time in a self-organised way. Also management structures, do you go through the structure or find your own path? Economists would say Game theory. but as the number of bodies grows it becomes more and more difficult to predict.
So large number interacting agents; hererogeneity; agents act to maximise their utility; feedback and repeated iterations.
Application example: e-mail perception of users is that cost of transport equals the cost of email which appears to be zero. So make the users aware of the real cost.
Say Cost of e-mail received is £3 but since the scale isn’t linear, congestion can be very expensive. So perhaps have dynamic pricing – cost of email set to reflect load on network. Make people aware of the price right now.
One Comment