I ran the durbinWatsonTest() in R and got p-value=0 for non time series data. Is this possible? I used durbinWatsonTest() on a non-time-series dataset: 
durbinWatsonTest(automobileRegression)

And the following is the result I get:
lag Autocorrelation D-W Statistic p-value
  1       0.3437134      1.309245       0
Alternative hypothesis: rho != 0

Is that possible to get the p value=0?  Any advice is appreciated.
 A: Short answer: yes, it's possible. 
Long answer: The Durbin Watson test is about the correlation of the residuals. If the data are ordered in some way, you'll get a significant DW test.
Here's an example:
> library(car)
> data(cars)
> fit1 <- lm(lm(speed ~ dist, data = cars))
> car::durbinWatsonTest(fit1)
 lag Autocorrelation D-W Statistic p-value
   1       0.3729415       1.19495   0.004
 Alternative hypothesis: rho != 0

My DW test is significant. 
Look at the data:
> head(cars, 12)
   speed dist
1      4    2
2      4   10
3      7    4
4      7   22
5      8   16
6      9   10
7     10   18
8     10   26
9     10   34
10    11   17
11    11   28
12    12   14
9     10   34

It appears that the data are sorted in order of speed. This makes the residuals autocorrelated and so the DW test is significant.
Let's put the data into a random order:
set.seed(1234)
cars <- cars[order(runif(nrow(cars))), ]

Fit the model, and do the test again:
fit2 <- lm(lm(speed ~ dist, data = cars))
car::durbinWatsonTest(fit2)

DW test is no longer significant. The test depends on the order of your data. 
