Urban Freight Management with Stochastic Time-Dependent Travel Times and Application to Large-Scale Transportation Networks

Table 2

Computational results and evaluation of routing plans in 4 delivery scenarios.

Instance specification			Vehicle number	Schedule time^b	Travel time	Failure number	Failure	CPU times (s)
Scenario	Number of Customers	Model					rate^c
1	25	TD	4	43.76	22.44	5.5	22%	5.55
		STD	5	47.35	23.61	0	0	5.27
			25.00%	8.20%	5.21%	None	−22%	−5%
2	50	TD	7	60.24	29.98	11.4	22.8%	19.63
		STD	9	65.62	31.20	0	0	17.08
			28.57%	8.93%	4.07%	None	−22.8%	−13%
3	100	TD	15	116.97	47.98	20.3	20.3%	71.11
		STD	17	126.89	53.25	0	0	66.58
			13.33%	8.48%	10.98%	None	−20.3%	−6%
4	150	TD	20	179.15	88.61	31.4	20.9%	148.38
		STD	23	202.99	98.08	0	0	137.88
			15.00%	13.31%	10.69%	None	−20.9%	−7%

indicates relative percentage change between the two tests.
^bTotal schedule time is the sum of total travel time (TT), total waiting time (WT), and total service time (SVT).
^cFailure rate = failure number/number of customers.