Rebalancing Docked Bicycle Sharing System with Approximate Dynamic Programming and Reinforcement Learning

Table 2

Key performance indicators using different benchmark strategies.

Strategy	Average unmet demand (person)	Average travel time (min)	Average delivery amount (bike)
No rebalance	10.6	—	—
STR	8.5	19.8	2.7
SLA	9.6	40.2	11.7
RTDP (1.00)	3.8	37.1	16.8
RTDP (1.65)	3.5	37.2	21.9
RTDP (2.33)	2.3	38.3	21.4