Notes

misc tricks

Always forget how to see the function definition of an S4 method:

 showMethods("xmlToS4", includeDefs=TRUE)

ropensci

  • Closed rfigshare issue #50
  • Still retooling reml to S4 and ReXML to inheritance-based S4 classes
  • building classes on the schema’s inheritance model 10:49 am 2013/08/26

In behavioral ecology we usually equate optimality with expected behavior arising from natural selection – i.e. the extent to which the observed behavior deviates from the optimal behavior is assumed to be due to modeling inaccuracies (optimizing the wrong thing).

In the operations research / engineering context, it seems optimization is more of a direct decision making tool to decide a strategy / design a controller, rather than to understand behavior that arises from the interplay of other forces.

I’m not quite clear what the standard interpretation is in the environmental economics / natural resource management context – do we also ascribe these deviations to modeling inaccuracy or rather to naive decision-makers (or a combination?)

I’m guessing we are closer to the behavioral ecologist’s viewpoint – that, for instance, the deviations observed in Figure 1 (what is “optimal” vs what policymakers do) can be explained primarily in terms of optimizing the wrong thing (ignoring the policy costs). I’m wondering if it will appear crazy or standard to assert that the observed behaviour truly is optimal relative to the true costs and constraints, and therefore, just as the ecologist would do, we might use the data to guide us towards understanding the nature of the correct constraints and costs.

A engineering minded critic might prefer a softer assumption – that the observed pattern reflects some design constraints perhaps (policy costs) but is not already “optimal” in reference to them, but rather some crude rule of thumb for lack of knowledge of how to determine the precise optimal tradeoff between adjustment costs and benefits. The difficulty with this latter approach is that it invites a third model for the ‘rule-of-thumb’ in addition to the optimal models with and without policy costs included.


ht @noamross for pointing out economists frequently assume deviations between predicted optimal and observed economic behavior (perhaps excluding policy..) are due to naiive models while we ecologists tend to assume they are the result of naiive policy.