Definition 12.3. A Boolean matrix is a real matrix whose every entry is either 0 or 1. Boolean entries 0/1 naturally model "off/on" in electrical switching circuits, or the adjacency matrix of a graph.
Given two Boolean matrices A=[aij] and B=[bij] of the same order, two new binary operations combine them entrywise:
Definition 12.4 (Join). A∨B=[aij∨bij]=[cij], where cij=1 if either aij=1 or bij=1, and cij=0 if both aij=0 and bij=0. Equivalently aij∨bij=max(aij,bij).
Definition 12.5 (Meet). A∧B=[aij∧bij]=[cij], where cij=1 only if both aij=1 and bij=1, and cij=0 if either entry is 0. Equivalently aij∧bij=min(aij,bij).
Worked pattern. For A=(0111), B=(1011): A∨B=(max(0,1)max(1,0)max(1,1)max(1,1))=(1111), and A∧B=(min(0,1)min(1,0)min(1,1)min(1,1))=(0011).
Which of the five properties do join and meet satisfy? Let B be the set of all Boolean matrices of a fixed order.
- Closure: since aij∨bij and aij∧bij are always 0 or 1, both operations are binary on B.
- Commutative: both hold, since max and min do not depend on the order of their two arguments.
- Associative: both hold, entrywise.
- Existence of identity: for ∨, the identity is the null matrix O (every entry 0) -- A∨O=O∨A=A. For ∧, the identity is the matrix U of all 1s -- A∧U=U∧A=A. …