A nice statement of a behavioural economics program is due to Godfrey Cadogan:
Is there a behavioral stochastic process peculiar to human behaviour? If so, what does it look like mathematically? So far, the Brownian motion or Levy process paradigm has been used to describe aspects of behavioral stochastic processes. How can we embed behavioural parameters in stochastic processes? (flakey link)
Things to think about here: Bounded rationality, rational inattention, institutions as stable orbits in behavioural systems...