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

Table 2

Comparison of the results between Lingo and SA_BRPSD (Q = 10 and C_h = 1).

Instance	Lingo		SABRPSD			GapBest (%)	GapAvg (%)
	Optimal value	CPU	Best value	Average value	CPU
Washington (20)1	668.94	3600	662.46	662.50	12.63	−0.97	−0.96
Washington (20)2	576.79	663	576.79	576.79	24.63	0.00	0.00
Chicago (20)1	164.14	3600	164.14	164.14	19.19	0.00	0.00
Chicago (20)2	136.83	3600	135.19	135.22	16.13	−1.20	−1.17
Washington (30)1	447.37	3600	393.75	393.75	34.19	−11.99	−11.99
Washington (30)2	501.14	3600	414.52	417.24	28.02	−17.28	−16.74
Chicago (30)1	300.48	3600	299.42	300.05	34.42	−0.35	−0.14
Chicago (30)2	299.76	3600	242.00	242.69	33.23	−19.27	−19.04
Washington (40)1	736.88	3600	688.63	692.12	60.31	−6.55	−6.07
Washington (40)2	911.88	3600	855.64	863.20	38.21	−6.17	−5.34
Chicago (40)1	576.94	3600	449.41	449.41	47.47	−22.10	−22.10
Chicago (40)2	538.80	3600	243.16	245.98	50.99	−54.87	−54.35
Washington (50)1			581.86	588.84	85.98
Washington (50)2			647.38	650.63	75.82
Chicago (50)			353.99	354.51	69.17
Washington (66)			676.54	683.93	90.98
Chicago (66)			6192.78	6468.38	120.21
Washington (80)1			733.58	744.35	114.69
Washington (80)2			736.77	745.27	183.02
Washington (90)			984.90	996.38	228.76
Avg		3355			68.40	−11.73	−11.49