smartgrid.rewards.numeric.differentiated.adaptability

Adaptability rewards change their definition as time goes by.

As their name indicates, they allow testing the agents’ capability to adapt to such changes: can their behaviour evolve with the new expectations?

These changes can be incremental, i.e., adding new objectives after some steps, or more brutal, i.e., completely replacing the targeted objectives by others.

members:

Classes

AdaptabilityOne()

Equity when t<3000, MultiObjectiveSum otherwise.

AdaptabilityThree()

Equity when t<2000, (Equity+OverConsumption)/2 when t<6000, (Equity+OC+Comfort)/3 otherwise.

AdaptabilityTwo()

Equity when t<2000, (Equity+OverConsumption)/2 otherwise.