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

Table 3

Comparison of the results between Lingo and SA_BRPSD (Q = 20 and C_h = 0).

Instance	Lingo		SABRPSD			GapBest (%)	GapAvg (%)
	Optimal value	CPU	Best value	Average value	CPU
Washington(20)1	645.10	3600	627.85	627.90	13.76	−2.67	−2.67
Washington(20)2	556.26	40	556.26	556.26	33.82	0.00	0.00
Chicago(20)1	134.73	3600	134.73	134.73	20.53	0.00	0.00
Chicago(20)2	100.80	3600	99.60	99.76	17.95	−1.19	−1.04
Washington(30)1	414.18	3600	369.72	369.72	32.90	−10.73	−10.73
Washington(30)2	432.25	3600	392.72	393.61	27.89	−9.15	−8.94
Chicago(30)1	489.71	3600	249.47	249.80	35.62	−49.06	−48.99
Chicago(30)2	321.25	3600	189.40	191.21	31.73	−41.04	−40.48
Washington(40)1	721.11	3600	617.41	622.76	50.49	−14.38	−13.64
Washington(40)2	1021.87	3600	814.77	819.79	42.10	−20.27	−19.78
Chicago(40)1	412.36	3600	408.94	408.94	57.98	−0.83	−0.83
Chicago(40)2	397.27	3600	201.52	203.18	50.88	−49.27	−48.86
Washington(50)1			526.11	534.88	109.26
Washington(50)2			592.72	595.78	79.45
Chicago(50)			321.13	324.60	69.87
Washington(66)			656.99	663.09	105.25
Chicago(66)			6243.75	6494.41	122.49
Washington(80)1			702.93	716.69	126.66
Washington(80)2			661.88	673.31	171.68
Washington(90)			914.40	924.30	204.03
Avg		3303.33			70.22	−16.55	−16.33