We use optional analytics, advertising and affiliate technologies to
measure usage, attribute advertising and improve EverythingQuant.
You can accept or reject non-essential tracking.
In a group of 11 people, a secure box is protected by multiple locks. The objective is to design a locking mechanism such that any group of 6 or more people can together open the box, but no group of fewer than 6 people is able to do so. Each person can carry a number of keys, where each key opens exactly one lock. A lock may be opened by multiple keys, held by different people. What is the minimum number of locks required to ensure this access control? Additionally, what is the minimum number of keys that each person must hold in order to satisfy the condition?
Solution
Let \(n=11\) people, and let the threshold be \(6\).
So we want:
\[
\text{any }6\text{ or more people can open the box,}
\]
but
\[
\text{no }5\text{ or fewer people can open it.}
\]
The standard way to think about this is to assign locks to the forbidden groups.
A group of fewer than \(6\) people is blocked if there is at least one lock for which none of them has a key.
Since any smaller group is contained in some \(5\)-person group, it is enough to make sure that every \(5\)-person group is blocked. If every \(5\)-person group fails, then every group of size \(<5\) also fails automatically.
So the real problem is:
\[
\text{for every }5\text{-person subset, there must be some lock that those }5\text{ cannot open.}
\]
Now take any lock. Suppose the people who do not have its key form a set \(S\).
If \(|S| \ge 6\), then those \(6\) people could not open the box, which is forbidden. Therefore every lock can be missing from at most \(5\) people.
That observation is the whole lower bound:
- every forbidden \(5\)-person group must be excluded from some lock;
- a lock can exclude at most one \(5\)-person group, because it can be missing from at most \(5\) people.
Hence we need at least as many locks as there are \(5\)-person groups:
\[
\binom{11}{5} = 462.
\]
This lower bound is achievable:
For each \(5\)-person subset \(S\), create one lock that is keyed to exactly the other \(6\) people, and not keyed to the people in \(S\).
Then:
- that particular \(5\)-person group \(S\) cannot open that lock;
- any \(6\)-person group can open every lock, because no \(6\)-person group is contained in a \(5\)-person set, so for every lock at least one of the \(6\) has a key.
So the minimum number of locks is
\[
\binom{11}{5}=462.
\]
Now count how many keys each person must hold.
Fix one person \(P\).
They receive a key for every lock whose forbidden \(5\)-set does not contain \(P\).
So we count the \(5\)-subsets chosen from the other \(10\) people:
\[
\binom{10}{5}=252.
\]
Thus each person must hold
\[
252
\]
keys.
You can also check this by averaging total keys:
each of the \(462\) locks is opened by exactly \(6\) people, so the total number of keys is
\[
462 \cdot 6 = 2772.
\]
Dividing equally among \(11\) people gives
\[
\frac{2772}{11}=252.
\]
Therefore the answers are
\[
\text{minimum number of locks} = 462,
\]
and
\[
\text{minimum number of keys per person} = 252.
\]