Affiliation:
1. Tsinghua University, Beijing, China
2. City University of Hong Kong, Hong Kong, China
Abstract
We prove a new discrete fixed point theorem for direction-preserving functions defined on integer points, based on a novel characterization of boundary conditions for the existence of fixed points. The theorem allows us to derive an improved algorithm for finding such a fixed point. We also develop a new lower bound proof technique. Together, they allow us to derive an asymptotic matching bound for the problem of finding a fixed point in a hypercube of any constantly bounded finite dimension.
Exploring a linkage with the approximation version of the continuous fixed point problem, we obtain asymptotic matching bounds for the complexity of the approximate Brouwer fixed point problem in the continuous case for Lipschitz functions. It settles a fifteen-years-old open problem of Hirsch, Papadimitriou, and Vavasis by improving both the upper and lower bounds.
Our characterization for the existence of a fixed point is also applicable to functions defined on nonconvex domains, which makes it a potentially useful tool for the design and analysis of algorithms for fixed points in general domains.
Funder
National Natural Science Foundation of China
Research Grants Council, University Grants Committee, Hong Kong
Ministry of Science and Technology of the People's Republic of China
Publisher
Association for Computing Machinery (ACM)
Subject
Artificial Intelligence,Hardware and Architecture,Information Systems,Control and Systems Engineering,Software
Cited by
17 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献