Adjacent Possible Innovation Dynamics on Local Optima Networks
arXiv preprint arXiv:2605.01821
Summary
We propose Local Optima Networks (LONs) as a model of how innovation unfolds. Each node is a locally stable technological configuration; weighted edges give the probability of moving from one basin of attraction to another. Innovators are stochastic walkers exploring the adjacent possible on this network.
A single model reproduces four regularities observed in discovery processes: sublinear growth of novelties, heavy-tailed frequency distributions, anomalous fluctuation scaling and power-law inter-event times. The exponents are constrained by the topology of the LON, and its communities offer an operational definition of technological paradigms.
