General Solution for Josephus Problem

YoavFrancis 7,469 views 14 slides May 13, 2015
Slide 1
Slide 1 of 14
Slide 1
1
Slide 2
2
Slide 3
3
Slide 4
4
Slide 5
5
Slide 6
6
Slide 7
7
Slide 8
8
Slide 9
9
Slide 10
10
Slide 11
11
Slide 12
12
Slide 13
13
Slide 14
14

About This Presentation

Overview of the solution for Josephus problem where every third person is eliminated, followed by a solution for the general case (arbitrary "q" where every q'th person is eliminated).

In addition a short discussion of interesting problem deriving from Josephus Problem.


Slide Content

A General Solution to Josephus Problem Yoav Francis July 2007 Topics in Number Theory, 2007/2 COMAS – College of Management – Academic Studies, Rishon Lezion, Israel Overview and Additional Problems

Motivation Given , find , where denotes the number of people in the initial circle. Let us first consider the case in which .  

Recursive Solution to   Where:  

A General Method for   For a group of people, each time we iterate on a person, we can provide him with a new number. In such a scenario – the numbers 1,2 would become . The number 3 would then be killed. 4,5 would become . 6 would be killed, and so on. Finally, would become , and would be killed and so on. Eventually the person denoted as will be the sole survivor.   Illustration for :  

A General Method for   The person to be killed is denoted (Example: the 10 th person is numbered “30”, the 9 th is numbered “27” and so on) We can find out who is the survivor if we find the original number of the person numbered (let us mark ) If then the person numbered necessarily had a previous number, which we can find in the following manner: or (Similarly to how we found the next number of and in the previous slide) Thus . The previous number of the person was or – meaning, it was (using )  

A General Method for   We have received an algorithm for :  

A General Method for   Let us note that this can be somewhat improved. The algorithm is not recursive, neither has a “closed” shape – but still allows us to compute the answer rather quickly for a large . Luckily, we can improve it. Let us define and use it instead of . This change matches numbering from to 1 instead of from to . This yields the following:  

A General Method for   We can now write the algorithm in the following manner: This is a much simpler form as enters the computation in a simple manner.  

The General Algorithm for Josephus Problem In a matter of fact, we can use the same justifications we used to show that the survivor for (where every person is killed) is computed in the following way:   This is the general algorithm. Let us note that for we’d get in every phase , so for the case of we get:  

The General Algorithm for Josephus Problem We can attempt to bring the algorithm to a “nicer” form. It cannot be brought to a closed form (but for the case of ), due to the “ceiling” function that we perform on every phase. However – we can still tidy-up the algorithm. Let us recursively define:   These numbers do not have a “nice” form but for , but if we agree to accept the expression as a “known expression” – then the solution for the general case of the Josephus problem would be:   (Where is the smallest one for which the following holds: )  

Additional Problems Derived From Josephus Problem Finding the second-to-last person remaining in Josephus problem (For example, if we want to find the last 2 people standing). Let us denote and get:  

Additional Problems Derived From Josephus Problem Another problem: Josephus finds himself in the position . Given that he can choose the “skip factor”( , which should he choose so he stays alive? Let us define and assume that . By Bertrand’s Postulate there exists a prime number . We can also assume that . We’d want to choose a q so the following holds: and:  

Additional Problems Derived From Josephus Problem Now, the people who would be removed (“killed”) are: The position would not be killed (we may get but that is still valid for the problem) Example: . Josephus got and will decide: (This holds for the 2 requirements of ). LCM(1,….,10) = 2520  

Thanks