Induction proof with handshakes

In summary, there are at least 2 people in a room, and each person in the room shakes hands with everyone else, but not with himself. The number of handshakes is (n^2-n)/2.
  • #1
nicnicman
136
0

Homework Statement


Suppose that if there are at least 2 people in a room and each person in the room shakes hands with everyone else, but not with himself. Show that the number of handshakes is (n^2-n)/2.
Make sure to show P(1), P(k) and prove P(k+1)



Homework Equations





The Attempt at a Solution


Well, I don't have too much yet, but here it is:

Let P(n) be 1 + 2 + 3 + ... + (n - 1) = (n^2-n)/2.

Now, I think I'm going the right direction but when I try to show P(1) I find it doesn't work, because (1^2-1)/2 = 0/2 = 0.

I'm missing something, but I don't know what it is.

Thanks for any suggestions.
 
Physics news on Phys.org
  • #2
How many ways there are for one person to shake hands with himself? Perhaps you should start with n=2.
 
  • #3
Zero I guess. I think it would be better if I began with 0.

Let P(n) be 0 + 1 + 2 + ... + (n -1) = (n^2-n)/2.

Now, (1^2-1)/2 = 0/2 = 0 and the sum of the first member in the list is 0.

Does that make sense?
 
  • #4
Your value of 0 in P(1) is quite correct. If one person is alone in the room and he shakes hands "with everyone else, but not with himself", no shaking will be done.

If you want to test the formula (which is correct, by the way), convince yourself by looking at n=2 people or n=3. Another way of getting a grasp on the situation modeled here is to draw n points on a piece of paper (standing for n people) and connecting each pair of points by a line (needn't be straight, it just stands for a handshake).

Draw the whole diagram for n=4 and do some counting. Then add a fifth point and add the necessary lines. You might get an idea about how to prove the step from P(n) to P(n+1).
 
  • #5
nicnicman said:
Zero I guess.
Correct. Does P(1)=0 satisfy P(N)=(N2-N)/2 for N=1? If it doesn't, the conjectured relationship is invalid (one way to disprove some supposedly universal conjecture is to find a specific counterexample for which the conjecture is false). On the other hand, if the conjecture is valid for N=1, you have a base case for the induction.

The other half of induction is the inductive step. Assume the relationship is valid for some integer k≥1 and show that this means that the relationship is also true for k+1.

So, suppose that there are some number k≥1 people in the room, all of whom have shaken hands with one another (but not themselves). Assume that the conjectured relationship P(N)=(N2-N)/2 is valid for N=k. Now Bob suppose enters the room and people start shaking hands again.
  1. How many more handshakes are needed to satisfy the condition that everyone in the room has shaken hands with everyone else, but not with himself?
  2. What is P(k+1) in terms of P(k) and the answer to question #1?
  3. Is the conjectured relationship P(N)=(N2-N)/2 valid for N=k+1?
 
  • #6
So, was I correct in beginning with 0 rather than 1. To me, it seems to work better.
 
  • #7
nicnicman said:
So, was I correct in beginning with 0 rather than 1. To me, it seems to work better.
No! You answered the question for N=1, not N=0. More importantly, the problem statement specifically said to start with N=1.
 
  • #8
D H said:
No! You answered the question for N=1, not N=0. More importantly, the problem statement specifically said to start with N=1.

Well, it also says that there are at least 2 people in the room.
 
  • #9
The problem statement says there are at least 2 people in the room, but it also tells you to start with P(1). This seems misleading, and I'm sure no one would complain if you include the cases

-- 1 person => 0 handshakes,
-- 1 handshake (2 people),

since either could be meant by "P(1)". For completeness' sake you could also consider the case of zero people, but I'd not go that far.
 

FAQ: Induction proof with handshakes

How does induction proof work for handshakes?

Induction proof is a mathematical technique used to prove a statement for all possible cases. In the case of handshakes, we use induction to prove that the number of handshakes in a group of n people is equal to n(n-1)/2. This means that the number of handshakes increases by 1 each time a new person joins the group, making it a recursive process.

Why is induction proof useful for handshakes?

Induction proof is useful for handshakes because it allows us to prove a statement for all possible cases without having to explicitly count each case. This is particularly helpful when dealing with large numbers or infinite cases.

Can induction proof be applied to other scenarios besides handshakes?

Yes, induction proof can be applied to various mathematical and logical statements. It is commonly used in algebra, geometry, and computer science to prove theorems, formulas, and algorithms.

What is the base case in an induction proof for handshakes?

The base case in an induction proof for handshakes is when there is only one person in the group. In this case, there are no handshakes, and the formula n(n-1)/2 still holds true, as 1(1-1)/2 = 0.

How do you prove the inductive step in an induction proof for handshakes?

The inductive step in an induction proof for handshakes involves assuming that the formula n(n-1)/2 is true for a group of n people, and then proving that it is also true for a group of n+1 people. This is usually done by showing that the number of handshakes increases by 1 when a new person joins the group, thus following the recursive nature of the formula.

Similar threads

Replies
15
Views
2K
Replies
7
Views
861
Replies
3
Views
842
Replies
11
Views
3K
Replies
4
Views
1K
Replies
3
Views
2K
Replies
6
Views
2K
Back
Top