Alternative objective functions to optimize

[hdrake]

One could imagine a number of different choices for the optimization scheme:

• minimizing total climate costs (damages + climate controls)
• minimizing damages given a fixed budget for climate controls
• minimizing control costs given a fixed temperature threshold
• minimize risk

[ron-rivest]

These are all good to think about!

Another one is to minimize the annual cost for a fixed policy that repeats every year but that stays within a given temperature bound.

[hdrake]

Good point @ron-rivest! Currently, costs vary dramatically from year to year, which is probably undesirable to policy makers. Some of this could be mitigated by adding latency in our economics (e.g. investments in control technologies in 2020 lead to emissions reductions (or other controls) that are effectively spread out over many years, possibly with some lag between the investment date and when the first controls go into effect). I believe your formulation has this qualitative behavior built in, while my formulation does not.

[hdrake]

I implemented the first three optimization options and merged to master in commit https://github.com/hdrake/OptimizeClimate/commit/7132e730aa8b86fda8d57005b36fbfaba30f59d7

I would need to come up with an appropriate metric to represent "risk" before I can optimize for that.