Address retirement mechanism for subnational regions

[robbieorvis]

The economic retirement mechanism retires capacity based on its profitability in MW/($/MW). If there are 3+ years of losses, the model multiplies the losses by the calibrated retirement parameter.

The retirement parameter is difficult to calibrate. For subnational regions, we tried downscaling based on a ratio of the state's capacity by resource to the national total. However, this doesn't really work.

For example, if a state has 1 GW of nuclear relative to the national value of 100 GW, then to retire 1 GW it would need to be 100x as unprofitable as plants nationally. However, simply using the national value also doesn't work, because it can mean that a slightly unprofitable resource type could dramatically over-retire.

We'll need to think of a workaround or better approach in the future to scale appropriately. It might require something like the ratio of that state's resource capacity divided by the ratio of the total in state capacity to national capacity.