zach401
commented
5 years ago Proposed Design
Design will require a new subclass of ChargingNetwork (ProbabilisticNetwork?) and one of EV (ProbabilisticEV?).
ProbabilisticEV
New variables
- affinity_map (dict) Dict mapping station_ids to a probability mass which represents the user's preferences over spaces.
- unplug_delay (int) Number of periods between when a vehicle finishes charging and when the driver moves their vehicle. Can use None or inf to represent that the driver will not leave until their departure time.
- unqueue_delay (int) Number of periods between when a driver is notified a spot is empty and when they actually plug in. Can use None or inf to represent that the driver is not willing to use the queue system.
ProbabilisticNetwork
New variables
-
waiting_queue (queue) - Queue holding EVSEs who are waiting for a spot to open. Not including those who have been notified.
-
notified (set) - Set of session_ids for user's who have been notified that they can exit the queue.
-
plugin()
- if EV session_id in notified or waiting_queue is empty and the number of open spaces > len(notified)
- Select an empty EVSE based on the EV's affinity_map.
- else
- Place EV in the queue
- if EV session_id in notified or waiting_queue is empty and the number of open spaces > len(notified)
-
unplug()
- If waiting_queue is not empty, remove the first EV from the queue and create a PluginEvent with the EV and timestamp t + unqueue_delay.
-
When an EV is done charging and its unplug_delay is < inf or not None, an UnplugEvent should be generated with timestamp t + unplug_delay. The original UnplugEvent should be deleted from the queue.
Feature
In some experiments it is helpful to not assign users to spaces deterministically, but rather according to some probability distribution over the open spaces in the network.
In the case that no spaces are available, the EV should be placed into a queue and wait for the next available space (possibly with a delay to model switching time).