Expected number of slots required for channel access - Birthday paradox
up vote
1
down vote
favorite
Let $n$ be the number of users who want to access a channel divided into with $s$ time-slots and let $n<<s$.
Each user randomly chooses one time-slot (out of $s$) for accessing the channel.
A collision occurs if more than one user chooses a given time-slot.
The probability of successful access for everybody without any collisions ( considering $n<<s$) can be approximated by $e^{-n^2/2s}$.
The expected number of users who experience a collision is also given by $n(1-(1-1/s)^{n-1})$.
This part of the problem is actually equivalent to the birthday problem where a year is assumed to have $s$ days.
Now, assume that in the case of a collision in the first round $i=1$ (time-slots $1,..., s$), the users who experience a collision will back off and try to access the channel, again randomly, during the next round $i = 2$ with $s$ time-slots (time-slots in $[s+1, 2s]$) and this procedure goes on during the next $s$ time-slots, if a collision occurs, until everybody accesses the channel.
The question is, how many rounds are required on average so that all the users access the channel once?
What would be a closed-form approximation?
probability expected-value birthday
add a comment |
up vote
1
down vote
favorite
Let $n$ be the number of users who want to access a channel divided into with $s$ time-slots and let $n<<s$.
Each user randomly chooses one time-slot (out of $s$) for accessing the channel.
A collision occurs if more than one user chooses a given time-slot.
The probability of successful access for everybody without any collisions ( considering $n<<s$) can be approximated by $e^{-n^2/2s}$.
The expected number of users who experience a collision is also given by $n(1-(1-1/s)^{n-1})$.
This part of the problem is actually equivalent to the birthday problem where a year is assumed to have $s$ days.
Now, assume that in the case of a collision in the first round $i=1$ (time-slots $1,..., s$), the users who experience a collision will back off and try to access the channel, again randomly, during the next round $i = 2$ with $s$ time-slots (time-slots in $[s+1, 2s]$) and this procedure goes on during the next $s$ time-slots, if a collision occurs, until everybody accesses the channel.
The question is, how many rounds are required on average so that all the users access the channel once?
What would be a closed-form approximation?
probability expected-value birthday
add a comment |
up vote
1
down vote
favorite
up vote
1
down vote
favorite
Let $n$ be the number of users who want to access a channel divided into with $s$ time-slots and let $n<<s$.
Each user randomly chooses one time-slot (out of $s$) for accessing the channel.
A collision occurs if more than one user chooses a given time-slot.
The probability of successful access for everybody without any collisions ( considering $n<<s$) can be approximated by $e^{-n^2/2s}$.
The expected number of users who experience a collision is also given by $n(1-(1-1/s)^{n-1})$.
This part of the problem is actually equivalent to the birthday problem where a year is assumed to have $s$ days.
Now, assume that in the case of a collision in the first round $i=1$ (time-slots $1,..., s$), the users who experience a collision will back off and try to access the channel, again randomly, during the next round $i = 2$ with $s$ time-slots (time-slots in $[s+1, 2s]$) and this procedure goes on during the next $s$ time-slots, if a collision occurs, until everybody accesses the channel.
The question is, how many rounds are required on average so that all the users access the channel once?
What would be a closed-form approximation?
probability expected-value birthday
Let $n$ be the number of users who want to access a channel divided into with $s$ time-slots and let $n<<s$.
Each user randomly chooses one time-slot (out of $s$) for accessing the channel.
A collision occurs if more than one user chooses a given time-slot.
The probability of successful access for everybody without any collisions ( considering $n<<s$) can be approximated by $e^{-n^2/2s}$.
The expected number of users who experience a collision is also given by $n(1-(1-1/s)^{n-1})$.
This part of the problem is actually equivalent to the birthday problem where a year is assumed to have $s$ days.
Now, assume that in the case of a collision in the first round $i=1$ (time-slots $1,..., s$), the users who experience a collision will back off and try to access the channel, again randomly, during the next round $i = 2$ with $s$ time-slots (time-slots in $[s+1, 2s]$) and this procedure goes on during the next $s$ time-slots, if a collision occurs, until everybody accesses the channel.
The question is, how many rounds are required on average so that all the users access the channel once?
What would be a closed-form approximation?
probability expected-value birthday
probability expected-value birthday
edited Nov 17 at 11:54
asked Nov 17 at 2:31
Mahdi
6910
6910
add a comment |
add a comment |
1 Answer
1
active
oldest
votes
up vote
1
down vote
accepted
The expected number of people who experience collisions in a round is
$$n(1 - (1 - 1/s)^{n-1}) approx n(1 - (1 - n/s)) = n^2 / s$$
as you noted. So we expect the number of people $P_i$ still participating after $i$ rounds to satisfy the recurrence
$$P_0 = n qquad P_{i+1} = frac{P_i^2}{s}$$
The first few values are $n, frac{n^2}{s}, frac{n^4}{s^3}, frac{n^8}{s^7}, cdots$ and the closed form is
$$P_i = frac{n^{2^i}}{s^{2^i-1}}$$
Solving for when this is less than 1,
$$n^{2^i} = s^{2^i - 1}$$
$$2^i log n = (2^i - 1)log s$$
$$log s = 2^i(log s - log n)$$
$$log log s = i + log(log s - log n)$$
$$i = log log s - log(log s - log n) = log(1 + frac{log n}{log s - log n})$$
which is approximately $log log n$ assuming $n << s$.
This is just a heuristic estimation of the expected number of rounds in total.
Great! Found something similar but not as clean as yours.
– Mahdi
Nov 17 at 11:54
Just two points. The number of rounds required at the end is $r=i+1$. Further, when I solve the problem for n less than 2 the approximation is much closer to the simulation. Perhaps because when we have less than two users, there would be no collision. I am not sure if this makes sense. I tried to approximate $(1-1/s)^{(n-1)}$ by $e^{(-n/s)}$ but could not find a closed form.
– Mahdi
Nov 20 at 9:14
add a comment |
1 Answer
1
active
oldest
votes
1 Answer
1
active
oldest
votes
active
oldest
votes
active
oldest
votes
up vote
1
down vote
accepted
The expected number of people who experience collisions in a round is
$$n(1 - (1 - 1/s)^{n-1}) approx n(1 - (1 - n/s)) = n^2 / s$$
as you noted. So we expect the number of people $P_i$ still participating after $i$ rounds to satisfy the recurrence
$$P_0 = n qquad P_{i+1} = frac{P_i^2}{s}$$
The first few values are $n, frac{n^2}{s}, frac{n^4}{s^3}, frac{n^8}{s^7}, cdots$ and the closed form is
$$P_i = frac{n^{2^i}}{s^{2^i-1}}$$
Solving for when this is less than 1,
$$n^{2^i} = s^{2^i - 1}$$
$$2^i log n = (2^i - 1)log s$$
$$log s = 2^i(log s - log n)$$
$$log log s = i + log(log s - log n)$$
$$i = log log s - log(log s - log n) = log(1 + frac{log n}{log s - log n})$$
which is approximately $log log n$ assuming $n << s$.
This is just a heuristic estimation of the expected number of rounds in total.
Great! Found something similar but not as clean as yours.
– Mahdi
Nov 17 at 11:54
Just two points. The number of rounds required at the end is $r=i+1$. Further, when I solve the problem for n less than 2 the approximation is much closer to the simulation. Perhaps because when we have less than two users, there would be no collision. I am not sure if this makes sense. I tried to approximate $(1-1/s)^{(n-1)}$ by $e^{(-n/s)}$ but could not find a closed form.
– Mahdi
Nov 20 at 9:14
add a comment |
up vote
1
down vote
accepted
The expected number of people who experience collisions in a round is
$$n(1 - (1 - 1/s)^{n-1}) approx n(1 - (1 - n/s)) = n^2 / s$$
as you noted. So we expect the number of people $P_i$ still participating after $i$ rounds to satisfy the recurrence
$$P_0 = n qquad P_{i+1} = frac{P_i^2}{s}$$
The first few values are $n, frac{n^2}{s}, frac{n^4}{s^3}, frac{n^8}{s^7}, cdots$ and the closed form is
$$P_i = frac{n^{2^i}}{s^{2^i-1}}$$
Solving for when this is less than 1,
$$n^{2^i} = s^{2^i - 1}$$
$$2^i log n = (2^i - 1)log s$$
$$log s = 2^i(log s - log n)$$
$$log log s = i + log(log s - log n)$$
$$i = log log s - log(log s - log n) = log(1 + frac{log n}{log s - log n})$$
which is approximately $log log n$ assuming $n << s$.
This is just a heuristic estimation of the expected number of rounds in total.
Great! Found something similar but not as clean as yours.
– Mahdi
Nov 17 at 11:54
Just two points. The number of rounds required at the end is $r=i+1$. Further, when I solve the problem for n less than 2 the approximation is much closer to the simulation. Perhaps because when we have less than two users, there would be no collision. I am not sure if this makes sense. I tried to approximate $(1-1/s)^{(n-1)}$ by $e^{(-n/s)}$ but could not find a closed form.
– Mahdi
Nov 20 at 9:14
add a comment |
up vote
1
down vote
accepted
up vote
1
down vote
accepted
The expected number of people who experience collisions in a round is
$$n(1 - (1 - 1/s)^{n-1}) approx n(1 - (1 - n/s)) = n^2 / s$$
as you noted. So we expect the number of people $P_i$ still participating after $i$ rounds to satisfy the recurrence
$$P_0 = n qquad P_{i+1} = frac{P_i^2}{s}$$
The first few values are $n, frac{n^2}{s}, frac{n^4}{s^3}, frac{n^8}{s^7}, cdots$ and the closed form is
$$P_i = frac{n^{2^i}}{s^{2^i-1}}$$
Solving for when this is less than 1,
$$n^{2^i} = s^{2^i - 1}$$
$$2^i log n = (2^i - 1)log s$$
$$log s = 2^i(log s - log n)$$
$$log log s = i + log(log s - log n)$$
$$i = log log s - log(log s - log n) = log(1 + frac{log n}{log s - log n})$$
which is approximately $log log n$ assuming $n << s$.
This is just a heuristic estimation of the expected number of rounds in total.
The expected number of people who experience collisions in a round is
$$n(1 - (1 - 1/s)^{n-1}) approx n(1 - (1 - n/s)) = n^2 / s$$
as you noted. So we expect the number of people $P_i$ still participating after $i$ rounds to satisfy the recurrence
$$P_0 = n qquad P_{i+1} = frac{P_i^2}{s}$$
The first few values are $n, frac{n^2}{s}, frac{n^4}{s^3}, frac{n^8}{s^7}, cdots$ and the closed form is
$$P_i = frac{n^{2^i}}{s^{2^i-1}}$$
Solving for when this is less than 1,
$$n^{2^i} = s^{2^i - 1}$$
$$2^i log n = (2^i - 1)log s$$
$$log s = 2^i(log s - log n)$$
$$log log s = i + log(log s - log n)$$
$$i = log log s - log(log s - log n) = log(1 + frac{log n}{log s - log n})$$
which is approximately $log log n$ assuming $n << s$.
This is just a heuristic estimation of the expected number of rounds in total.
answered Nov 17 at 4:12
Chris Jones
787514
787514
Great! Found something similar but not as clean as yours.
– Mahdi
Nov 17 at 11:54
Just two points. The number of rounds required at the end is $r=i+1$. Further, when I solve the problem for n less than 2 the approximation is much closer to the simulation. Perhaps because when we have less than two users, there would be no collision. I am not sure if this makes sense. I tried to approximate $(1-1/s)^{(n-1)}$ by $e^{(-n/s)}$ but could not find a closed form.
– Mahdi
Nov 20 at 9:14
add a comment |
Great! Found something similar but not as clean as yours.
– Mahdi
Nov 17 at 11:54
Just two points. The number of rounds required at the end is $r=i+1$. Further, when I solve the problem for n less than 2 the approximation is much closer to the simulation. Perhaps because when we have less than two users, there would be no collision. I am not sure if this makes sense. I tried to approximate $(1-1/s)^{(n-1)}$ by $e^{(-n/s)}$ but could not find a closed form.
– Mahdi
Nov 20 at 9:14
Great! Found something similar but not as clean as yours.
– Mahdi
Nov 17 at 11:54
Great! Found something similar but not as clean as yours.
– Mahdi
Nov 17 at 11:54
Just two points. The number of rounds required at the end is $r=i+1$. Further, when I solve the problem for n less than 2 the approximation is much closer to the simulation. Perhaps because when we have less than two users, there would be no collision. I am not sure if this makes sense. I tried to approximate $(1-1/s)^{(n-1)}$ by $e^{(-n/s)}$ but could not find a closed form.
– Mahdi
Nov 20 at 9:14
Just two points. The number of rounds required at the end is $r=i+1$. Further, when I solve the problem for n less than 2 the approximation is much closer to the simulation. Perhaps because when we have less than two users, there would be no collision. I am not sure if this makes sense. I tried to approximate $(1-1/s)^{(n-1)}$ by $e^{(-n/s)}$ but could not find a closed form.
– Mahdi
Nov 20 at 9:14
add a comment |
Thanks for contributing an answer to Mathematics Stack Exchange!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
Use MathJax to format equations. MathJax reference.
To learn more, see our tips on writing great answers.
Some of your past answers have not been well-received, and you're in danger of being blocked from answering.
Please pay close attention to the following guidance:
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3001903%2fexpected-number-of-slots-required-for-channel-access-birthday-paradox%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown