Normal distribution vs.exponential

In summary, there is a question about why some articles assume a normal distribution for bus headway times, even though exponential distributions are typically used for modeling inter-arrival times. The reasons for this assumption may include the central limit theorem and information theory. However, it is unclear if this assumption is appropriate for modeling bus travel times.
  • #1
Mark J.
81
0
Sorry because I have asked this other times but still not getting a reasonable answer:

For bus headway times why in several articles they make assumption of normal distribution??
Generally inter-arrival times are modeled using exponential distribution but in some papers
http://onlinelibrary.wiley.com/doi/10.1002/atr.5670250304/abstract
http://www.sciencedirect.com/science/article/pii/0191261586900366

normal distribution is assumed.
Which are the reasons to make this assumption?
Maybe because the real time is calculated including time for passengers to enter in bus and departing?
Maybe because of normal distribution quality of modeling deviations from some point (scheduled time) ?

Please help
Regards
 
Mathematics news on Phys.org
  • #2
Exponential distributions are a family of distributions that include the normal distribution. Generally, whenever something can be modeled by an exponential distribution and there is no addition information to clearly specify which specific exponential distribution is to be used, a normal distribution will be assumed for various reasons that include the central limit theorem and some basic results from information theory (the normal distribution has the greatest amount of "uncertainty" associated with it out of all exponential distributions, and so, in a sense, it makes fewer assumptions than the others).
 
  • #3
Yes but it is rather vague.
I was thinking if travel time = bus headway + time spent at the stop then if bus headways for example are exponential and time spent at the stop some other distribution than I gues their sum is not exponential but someway renewal process so maybe there will be some approximation to normal?
 
  • #4
Number nine, the normal distribution is not an example of an exponential distribution.

Mark, reading the second paper (pdf found here http://www-bcf.usc.edu/~maged/publications/Optimal%20Holding%20Times.pdf ) I see that the amount by which a bus is late is modeled as a normal distribution, but I don't see where they talk about the amount of time between buses being a normal distribution (which doesn't really make sense because the distribution should have only non-negative numbers as its support)
 
  • #5
Office Shredder you are totally right.
Anyway is there any orientation about normal assumption in bus travel time modeling you can give please?
Best regards
 
  • #6
Number nine, the normal distribution is not an example of an exponential distribution.

I interpreted the OP (perhaps incorrectly) as being about something from the exponential family of distributions, of which the normal distribution is a member. In retrospect, that's almost certainly not what he's talking about.
 

FAQ: Normal distribution vs.exponential

What is the difference between normal distribution and exponential distribution?

Normal distribution is a probability distribution that is symmetrical and bell-shaped, with the majority of data points falling close to the mean. Exponential distribution, on the other hand, is a probability distribution that is skewed and has a long tail on one side, with data points declining rapidly as they move away from the mean.

How are the shapes of normal and exponential distributions different?

The shape of a normal distribution is symmetrical, with the mean, median, and mode all being equal. The shape of an exponential distribution, however, is skewed, with the mean being larger than the median and mode.

What types of data are commonly modeled by normal and exponential distributions?

Normal distribution is commonly used to model continuous data that follows a bell-shaped curve, such as heights, weights, and test scores. Exponential distribution is used to model data that is skewed and has a long tail, such as waiting times, failure rates, and the amount of time between events.

How are probabilities calculated for normal and exponential distributions?

For normal distribution, probabilities are calculated using the standard normal distribution table or by using a statistical software. For exponential distribution, probabilities are calculated using the exponential distribution formula or by using a statistical software.

What is the central limit theorem and how does it relate to normal and exponential distributions?

The central limit theorem states that as the sample size increases, the distribution of the sample means will approach a normal distribution, regardless of the shape of the population distribution. This means that even if the population follows an exponential distribution, the sample means will follow a normal distribution as the sample size increases.

Similar threads

Back
Top