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

Table 1

Comparison of the results between Lingo and SA_BRPSD (Q = 20 and C_h = 1).

Instance	Lingo		SABRPSD			GapBest (%)	GapAvg (%)
	Optimal value	CPU	Best value	Average value	CPU
Washington (20)1	639.15	3600	639.16	639.17	10.15	0.00	0.00
Washington (20)2	576.79	159	576.79	576.79	24.58	0.00	0.00
Chicago (20)1	166.24	3600	153.40	153.40	20.85	−7.72	−7.72
Chicago (20)2	115.19	3600	113.46	113.46	14.04	−1.50	−1.50
Washington (30)1	450.19	3600	387.78	388.03	30.77	−13.86	−13.81
Washington (30)2	520.39	3600	403.12	405.09	25.22	−22.53	−22.16
Chicago (30)1	518.02	3600	265.71	266.96	54.21	−48.71	−48.47
Chicago (30)2	204.14	3600	202.21	204.14	29.48	−0.94	0.00
Washington (40)1	797.87	3600	634.07	638.52	44.71	−20.53	−19.97
Washington (40)2	1036.51	3600	826.68	831.62	29.94	−20.24	−19.77
Chicago (40)1	1287.45	3600	424.34	424.34	52.13	−67.04	−67.04
Chicago (40)2	478.01	3600	213.10	215.92	48.37	−55.42	−54.83
Washington (50)1			546.18	554.30	105.62
Washington (50)2			603.40	611.72	77.76
Chicago (50)			326.32	328.59	70.84
Washington (66)			682.51	685.62	101.87
Chicago (66)			6243.75	6494.41	122.49
Washington (80)1			719.65	732.09	121.20
Washington (80)2			678.58	684.66	159.47
Washington (90)			925.21	938.56	196.93
Avg		3335			66.14	−24.15	−23.85