Factored Customers - Pyosch/powertac-server GitHub Wiki

up

The following are the factored-customer model instantiations included in the standard distribution. This file is derived from the documentation in the XML config file for the factored-customer model. Please refer to the config files for the latest details: https://github.com/powertac/factored-customer/blob/master/src/main/resources/config/.

BrooksideHomes

A prototypical multicontracting consumer population of 30000 households in a suburban area with consumption peaking in the morning and evening. The aggregate capacity is generated using a timeseries forecasting model learned from a fine-grained model based on the MEREGIO project household model developed by Gottwalt, et al.

CentervilleHomes

A multicontracting consumer population of 20000 households that is similar to BrooksideHomes in many aspects. There are three key distinctions: (i) the capacities for these customers are generated using a population distribution model instead of a timeseries forecastign model, (ii) the capacities exhibit decision-theoretic adaptive behavior whereby they adjust the time-shifting of capacities dynamically when presented with TOU tariffs, and (iii) tariffs are reevaluated more frequently.

DowntownOffices

A multicontracting consumer population of 30 urban offices that are similar to household populations in most aspects except that the consumption pattern is consistently sustained at a higher level during weekdays and is lower at night and on the weekends.

EastsideOffices

A multicontracting consumer population of 25 suburban offices similar to DowntownOffices in many aspects except that the generated capacities employ dynamically learned time-shifting behavior similar to the capabilities of CentervilleHomes and tariffs are reevaluated more frequently.

FrostyStorage

A centrally-managed set of 3 cold-storage facilities with interruptible consumption. These facilitiesshre a set of 3 capacity generators, each of which draws from an individual probability distribution. In the standard configuration, these generators are identical. The consumption capacity is somewhat lower at night than during the day, but does not vary by day of week. Interrupted capacity is shifted over to the next three timeslots. Since the population units are centrally managed, they use similar tariff evaluation criteria but the facilities are individually metered so they can be allocated to different tariffs if needed, i.e., the aggregate customer is multicontracting.

HextraChemical

A multicontracting installation of 5 chemical manufacturing plants with interruptible consumption. The capacity generated from each plant is drawn from a uniform distribution for all weekdays with minimal consumption over the weekends. Interrupted capacity is typically shifted over to the next four timeslots.

MedicalCenter

A hybrid customer representing a single large hospital complex with a large consumption capacity and a small solar production capacity. The consumption and production capacities may be allocated to tariffs from different brokers. Consumption is skewed towards slightly lower consumption at night and notably lower over the weekend, although much higher than is typical for commercial office buildings. Production capacity is mostly governed by daylight and cloud cover.

SunnyhillSolar

A small community-owned PEV farm that is managed as a single unit and therefore has no multicontracting capabilities. The generated capacity is driven largely by daylight and cloud cover.

WindmillCoOp

A cooperative of 90 rural wind turbines with muticontracting production. 50 of the wind turbines are placed to maximize production from south-westerly winds and 40 are placed to maximize production from south-easterly winds.