Logical Zonotope: A Set Representation for Binary Vectors
share
In this paper, we propose a new set representation for binary vectors called logical zonotopes. A logical zonotope is constructed by XOR-ing a binary vector with a combination of binary vectors called generators. A logical zonotope can efficiently represent up to 2^n binary vectors using only n generators. Instead of the explicit enumeration of the zonotopes' members, logical operations over sets of binary vectors are applied directly to a zonotopes' generators. Thus, logical zonotopes can be used to greatly reduce the computational complexity of a variety of operations over sets of binary vectors, including logical operations (e.g. XOR, NAND, AND, OR) and semi-tensor products. Additionally, we show that, similar to the role classical zonotopes play for formally verifying dynamical systems defined over real vector spaces, logical zonotopes can be used to efficiently analyze the forward reachability of dynamical systems defined over binary vector spaces (e.g. logical circuits or Boolean networks). To showcase the utility of logical zonotopes, we illustrate three use cases: (1) discovering the key of a linear-feedback shift register with a linear time complexity, (2) verifying the safety of a logical vehicle intersection crossing protocol, and (3) performing reachability analysis for a high-dimensional Boolean function.
READ FULL TEXT