We are seeing occasional cost estimates of 0.0 from the household and officecomplex customer models when evaluating time-of-use tariffs. This does not appear to happen immediately in a game, but rather after several evaluations. Once it starts happening, it seems to be repeated on other tariffs. It is observable in games 115 and 155 of the 2014 Qualifying rounds.
Here's an example from the trace log in game 155:
4294511 INFO TariffMarketService: publishing spec 300002584 broker: Maxon, exp: null
4294512 INFO common.TariffEvaluator: Evaluated tariff 300002584: cost=-8.0852E-319, inconvenience=0.05
4294512 INFO common.TariffEvaluator: Evaluated tariff 300002584: cost=0.0, inconvenience=0.05
4294513 INFO common.TariffEvaluator: Evaluated tariff 300002584: cost=-76.5948581924141, inconvenience=0.05
We are seeing occasional cost estimates of 0.0 from the household and officecomplex customer models when evaluating time-of-use tariffs. This does not appear to happen immediately in a game, but rather after several evaluations. Once it starts happening, it seems to be repeated on other tariffs. It is observable in games 115 and 155 of the 2014 Qualifying rounds.
Here's an example from the trace log in game 155: