A Geo-Aware and VRP-Based Public Bicycle Redistribution System

Table 4

Simulation results for the VRP of Washington Capital Bikeshare.

The simulation case: Washington Capital Bikeshare (41/132 station for service)
Parameters	Working time: 6 h
	The full capacity of vehicles: 30 bikes
	Prefetch: 10 bikes
	The average time of redistributing one bike: 3 min
	Station safety stock: upper limit (80%), lower limit (20%), adjust value (50%)
		Duration (h:min)	Route length (km)	Transit time (min)
The Nearest Neighbor Algorithm for the VRP (The Euclidean Distance)	Truck no. 1	6:07	38.243	82
	Truck no. 2	5:58	51.908	10
	Truck no. 3	5:49	38.344	86
	Truck no. 4	1:39	30.977	45
	Total distance: 159.472 km		Total duration: 19:33
The Nearest Neighbor Algorithm for the VRP (The Actual Distance Path)	Truck no. 1	5:58	34.650	83
	Truck no. 2	5:54	36.470	78
	Truck no. 3	5:55	45.628	94.6
	Truck no. 4	1:40	31.340	47
	Total distance: 148.088 km		Total duration: 19:27
The Savings Method for the VRP (The Euclidean Distance)	Truck no. 1	7:33	95.938	195
	Truck no. 2	7:03	72.764	147
	Truck no. 3	7:18	69.341	144
	Truck no. 4	0:43	10.601	14
	Total distance: 248.644 km		Total duration: 22:37
The Savings Method for the VRP (The Actual Distance Path)	Truck no. 1	5:31	61.545	121
	Truck no. 2	5:37	45.338	95
	Truck no. 3	5:18	46.072	102
	Truck no. 4	3:55	31.463	68
	Total distance: 184.418 km		Total duration: 20:21