Title :
The exact information complexity of Indian poker
Author :
Ozarow, Lawrence H. ; Shepp, L.A.
Author_Institution :
AT&T Bell Lab., Murray Hill, NJ, USA
fDate :
1/1/1990 12:00:00 AM
Abstract :
Suppose N people are in a circle and each has a K-bit integer on his forehead and can see the others´ integers but does not know his own. The problem is to determine how many bits of information must be exchanged among the players so that each can determine his own number after the exchange. The authors provide an algorithm for which the number of bits that suffices equals K plus the least integer ⩾K/(N-1). In particular, if N>K, the total of NK bits can be resolved in K+1 transmissions. Since K bits are required to resolve any of the numbers, this is quite efficient, and is shown to be optimal. In addition, the optimal strategy is obtained when the numbers are of different lengths
Keywords :
information theory; protocols; Indian poker; exact information complexity; protocols; Forehead; IEEE news; Mirrors; Notice of Violation; Protocols; Uncertainty;
Journal_Title :
Information Theory, IEEE Transactions on