Author |
Topic: Take 1000 Odd Digits, Get Number (Read 466 times) |
|
K Sengupta
Senior Riddler
Gender:
Posts: 371
|
|
Take 1000 Odd Digits, Get Number
« on: Jul 24th, 2007, 7:24am » |
Quote Modify
|
Analytically determine the total number of 1000 digit positive decimal integers constituted entirely by odd digits with the proviso that any two adjacent digits must differ by 2.
|
|
IP Logged |
|
|
|
Barukh
Uberpuzzler
Gender:
Posts: 2276
|
|
Re: Take 1000 Odd Digits, Get Number
« Reply #1 on: Jul 24th, 2007, 11:50pm » |
Quote Modify
|
3501 - 3499? Analysis: hidden: | Let Tn be the number of n-digit numbers having the required property, and let Tn(d) be the number of such numbers starting with digit d. Then we have the following relations: 1) Tn = 2Tn(1) + 2Tn(3) + Tn(5) 2) Tn(1) = Tn-1(3) 3) Tn(3) = Tn-1(1) + Tn-1(5) 4) Tn(5) = 2Tn-1(3) 5) T1(1) = T1(3) = T1(5) = 1 By induction, it can be shown that T2n(1) = 3n-1, T2n(3) = T2n(5) = 2T2n(1). |
|
« Last Edit: Jul 24th, 2007, 11:51pm by Barukh » |
IP Logged |
|
|
|
jollytall
Senior Riddler
Gender:
Posts: 585
|
|
Re: Take 1000 Odd Digits, Get Number
« Reply #2 on: Jul 24th, 2007, 11:55pm » |
Quote Modify
|
Two different odd numbers always differ by 2 at least. So probably you wanted to say something else. This way it is really easy (doesn't worth even hiding). The last digit can be anything (i.e. 5 choices). The next can be 4 different numbers, and so on. So I would go for 5*4^999. What do I miss? Did you want to say "differ by 4"?
|
|
IP Logged |
|
|
|
Barukh
Uberpuzzler
Gender:
Posts: 2276
|
|
Re: Take 1000 Odd Digits, Get Number
« Reply #3 on: Jul 25th, 2007, 12:35am » |
Quote Modify
|
on Jul 24th, 2007, 11:55pm, jollytall wrote:Two different odd numbers always differ by 2 at least. |
| I think it says that the adjacent digits differ by exactly 2. So, for instance, if the first digit of the number is 1, you've got just one choice for the second (3), while if it starts from 3, you have two choices (1, 5).
|
|
IP Logged |
|
|
|
towr
wu::riddles Moderator Uberpuzzler
Some people are average, some are just mean.
Gender:
Posts: 13730
|
|
Re: Take 1000 Odd Digits, Get Number
« Reply #4 on: Jul 25th, 2007, 12:36am » |
Quote Modify
|
on Jul 24th, 2007, 11:55pm, jollytall wrote:Two different odd numbers always differ by 2 at least. So probably you wanted to say something else. This way it is really easy (doesn't worth even hiding). The last digit can be anything (i.e. 5 choices). The next can be 4 different numbers, and so on. So I would go for 5*4^999. What do I miss? Did you want to say "differ by 4"? |
| He might mean differ by 2 exactly, rather than at least 2. in which case for 1 or 9, the next digit can be only 3 or 7 respectively, while in other cases you have two choices.
|
|
IP Logged |
Wikipedia, Google, Mathworld, Integer sequence DB
|
|
|
K Sengupta
Senior Riddler
Gender:
Posts: 371
|
|
Re: Take 1000 Odd Digits, Get Number
« Reply #5 on: Jul 25th, 2007, 2:06am » |
Quote Modify
|
on Jul 24th, 2007, 11:55pm, jollytall wrote:Two different odd numbers always differ by 2 at least. ........... Did you want to say "differ by 4"? |
| I would like to clarify that any two adjacent digits of each of the numbers must have a difference of precisely 2.
|
« Last Edit: Jul 25th, 2007, 2:08am by K Sengupta » |
IP Logged |
|
|
|
K Sengupta
Senior Riddler
Gender:
Posts: 371
|
|
Re: Take 1000 Odd Digits, Get Number
« Reply #6 on: Jul 25th, 2007, 2:30am » |
Quote Modify
|
on Jul 24th, 2007, 11:50pm, Barukh wrote:3501 - 3499? Analysis: hidden: | Let Tn be the number of n-digit numbers having the required property, and let Tn(d) be the number of such numbers starting with digit d. Then we have the following relations: 1) Tn = 2Tn(1) + 2Tn(3) + Tn(5) 2) Tn(1) = Tn-1(3) 3) Tn(3) = Tn-1(1) + Tn-1(5) 4) Tn(5) = 2Tn-1(3) 5) T1(1) = T1(3) = T1(5) = 1 By induction, it can be shown that T2n(1) = 3n-1, T2n(3) = T2n(5) = 2T2n(1). | |
| True. My independent computations also show that 8*3499 indeed is the required answer.
|
« Last Edit: Jul 25th, 2007, 2:33am by K Sengupta » |
IP Logged |
|
|
|
Eigenray
wu::riddles Moderator Uberpuzzler
Gender:
Posts: 1948
|
|
Re: Take 1000 Odd Digits, Get Number
« Reply #7 on: Jul 25th, 2007, 4:34am » |
Quote Modify
|
And for an odd number n of digits, 14*3(n-3)/2.
|
|
IP Logged |
|
|
|
K Sengupta
Senior Riddler
Gender:
Posts: 371
|
|
Re: Take 1000 Odd Digits, Get Number
« Reply #8 on: Jul 26th, 2007, 7:46am » |
Quote Modify
|
I furnish my methodology leading to the end result of 8*3^499, which is fundamentally similar to the solution posted by Barukh. Solution: Let us denote: E_n = the number of such n-digit numbers with last digit 1 or 9, and: F_n = the number of such n- digit numbers with last digit 3 or 7 G_n = the number of such n-digit numbers with last digit 5 Then, we must have: E_(n+1) = F_n; F_(n+1) = 2* G_n + E_n, and: G_(n+1) = F_n Accordingly F_(n+2) = 2*G_(n+1) + E_(n+1) = 2*F_n + F_n = 3*F_n Evidently, F_1 = 2(3, 7), and F_2 = 4(13, 53, 57, 97) Accordingly, we must have: F_(2n) = 4*3^(n-1); F_(2n-1) = 2*3^(n-1). Therefore, E_(2n) = G_(2n) = 2*3^(n-1), and consequently: E_(2n) + F_(2n) + G_(2n) = 8*3^(n-1) Substituting n= 500, we obtain 8*3^499 as the desired result.
|
|
IP Logged |
|
|
|
K Sengupta
Senior Riddler
Gender:
Posts: 371
|
|
Re: Take 1000 Odd Digits, Get Number
« Reply #9 on: Jul 26th, 2007, 7:48am » |
Quote Modify
|
on Jul 25th, 2007, 4:34am, Eigenray wrote:And for an odd number n of digits, 14*3(n-3)/2. |
| I have been unable to work out a satisfactory methodology for the general case corresponding to an odd number n of digits, and consequently I am still looking for the foregoing procedure leading to the final solution.
|
|
IP Logged |
|
|
|
towr
wu::riddles Moderator Uberpuzzler
Some people are average, some are just mean.
Gender:
Posts: 13730
|
|
Re: Take 1000 Odd Digits, Get Number
« Reply #10 on: Jul 26th, 2007, 8:22am » |
Quote Modify
|
on Jul 26th, 2007, 7:48am, K Sengupta wrote:I have been unable to work out a satisfactory methodology for the general case corresponding to an odd number n of digits, and consequently I am still looking for the foregoing procedure leading to the final solution. |
| You can easily derive it from Barukh's solution. T2n+1(1) = T2n(3) = 2 T2n(1) T2n+1(3) = T2n(1) + T2n(5) = 3 T2n(1) T2n+1(5) = 2Tn-1(3) = 4 T2n(1) 2*2T2n(1)+2*3T2n(1)+4T2n(1) = 14*3n-1, or substituting in k=2n+1, 14*3(k-3)/2
|
« Last Edit: Jul 26th, 2007, 8:24am by towr » |
IP Logged |
Wikipedia, Google, Mathworld, Integer sequence DB
|
|
|
|