Addressing the Bike Repositioning Problem in Bike Sharing System: A Two-Stage Stochastic Programming Model

Table 4

Comparison of the results between Lingo and SA_BRPSD (Q = 10 and C_h = 0).

Instance	Lingo		SABRPSD			GapBest (%)	GapAvg (%)
	Optimal value	CPU	Best value	Average value	CPU
Washington(20)1	665.08	3600	654.91	655.01	10.76	−1.53	−1.52
Washington(20)2	566.26	30	566.26	566.26	25.00	0.00	0.00
Chicago(20)1	155.17	3600	155.17	155.17	19.20	0.00	0.00
Chicago(20)2	127.97	3600	127.97	127.97	17.38	0.00	0.00
Washington(30)1	449.66	3600	382.56	382.56	33.10	−14.92	−14.92
Washington(30)2	442.69	3600	410.45	411.18	28.40	−7.28	−7.12
Chicago(30)1	292.28	3600	289.57	290.67	33.25	−0.93	−0.55
Chicago(30)2	236.60	3600	234.10	234.10	32.92	−1.06	−1.06
Washington(40)1	818.25	3600	681.61	685.40	61.54	−16.70	−16.24
Washington(40)2	859.04	3600	853.74	857.84	47.10	−0.62	−0.14
Chicago(40)1	500.44	3600	442.74	442.74	47.85	−11.53	−11.53
Chicago(40)2	349.26	3600	239.51	240.61	53.41	−31.42	−31.11
Washington(50)1			571.78	579.44	106.77
Washington(50)2			638.40	640.92	75.32
Chicago(50)			352.67	354.72	72.96
Washington(66)			664.31	671.87	89.90
Chicago(66)			6175.14	6472.54	117.22
Washington(80)1			728.64	739.63	114.94
Washington(80)2			722.88	743.14	183.62
Washington(90)			984.90	996.38	228.76
Avg		3303			69.97	−7.17	−7.02