Measuring Service Reliability Using Automatic Vehicle Location Data
Table 1
Hypothesis tests for potential travel time distributions.
Single models
K-S test (-value)*
A-D test (-value)*
Parameters#
Goodness-of-fit AIC
Shape
Scale
Burr
0.111 (0.089)
1.574 (0.160)
15.86/1.59
30.00
−590
Exponential
0.563 (<0.05)
46.657 (<0.05)
28.89
N/A
−1071
Extreme value
0.107 (0.110)
2.160 (0.075)
30.25
2.75
−612
Gamma
0.111 (0.088)
1.364 (0.212)
118.19
0.24
−585
Log-normal
0.110 (0.097)
1.339 (0.220)
3.36
0.09
−584
Logistic
0.110 (0.096)
1.643 (0.146)
28.83
1.59
−595
Log-logistic
0.111 (0.091)
1.625 (0.149)
3.359
0.055
−593
Normal
0.112 (0.087)
1.372 (0.210)
28.89
2.68
−587
Weibull
0.091 (0.241)
1.679 (0.139)
11.17
30.13
−602
* value < 0.05 rejects the null hypothesis that the data come from the distribution.
#The scale parameter indicates the degree of the spread for travel time distribution. The shape parameter indicates the shape and location of the travel time distribution.