Self-Adaptive Algorithms for the Split Common Fixed Point Problem of the Demimetric Mappings

Xinhong Chen, Yanlai Song, He Jian-ying, Liping GongJournal of Applied Mathematics and Physics 2019 · DOI

Abstract

The split common fixed point problem (SCFPP) is an inverse problem that consists in finding an element in a fixed point set such that its image under a bounded linear operator belongs to another fixed-point set. In this paper, we present new iterative algorithms for solving the SCFPP of demimetric mappings in Hilbert spaces. Notably, our algorithms do not require any prior information regarding the operator norm. Weak and strong convergence theorems are established under mild assumptions, extending and improving upon recent results in the literature.

Introduction

The Split Common Fixed Point Problem (SCFPP) is a sophisticated inverse problem with foundational applications in signal processing, medical image reconstruction (such as CT scans), and intensity-modulated radiation therapy. Mathematically, the problem involves finding a point $x$ in a Hilbert space $H_1$ that is a fixed point of a mapping $S$ ($x \in Fix(S)$), such that its image under a bounded linear operator $A$ is a fixed point of another mapping $T$ in Hilbert space $H_2$ ($Ax \in Fix(T)$).

A significant computational hurdle in existing iterative solutions is the requirement for the operator norm $\|A\|$ to determine appropriate step sizes. In many real-world applications, calculating this norm is extremely difficult or computationally expensive. This paper addresses this challenge by introducing self-adaptive iterative algorithms that eliminate the need for prior knowledge of the operator norm while expanding the mathematical scope to a broader class of mappings.

Methodology and Mathematical Scope

The authors focus on demimetric mappings, a broad category that serves as a "super-set" for directed mappings, demicontractive mappings, and quasi-non-expansive mappings. By proving results for this general class, the paper unifies several sub-fields of fixed-point theory.

Self-Adaptive Step Sizes

The core innovation of the proposed methods is the use of a self-adaptive step size $\tau_n$. Instead of relying on a global constant derived from the operator norm, the algorithm calculates the step size dynamically based on the current state of the iteration. This "norm-free" approach allows the algorithm to function in environments where the global properties of the operator $A$ are unknown.

Iterative Algorithms

Two primary iterative schemes are presented:

  1. Algorithm 3.1 (Weak Convergence): This algorithm ensures that the sequence of guesses approaches the solution set in the weak topology of the Hilbert space. This is formalized in Theorem 3.2.
  2. Algorithm 3.3 (Strong Convergence): By incorporating a contraction mapping, this algorithm ensures the sequence converges strongly to a specific, unique solution point—specifically, the projection of the starting point onto the solution set. This is formalized in Theorem 3.4.

Key Insights

  • The "Norm-Free" Advantage: The removal of the operator norm requirement allows these algorithms to be applied to much larger or more complex operators where the norm cannot be easily estimated.
  • Demiclosedness at Zero: A critical technical assumption used is that the mappings are "demiclosed at zero." This property ensures that if a sequence of points appears to approach a fixed point, the limit of that sequence is indeed a fixed point.
  • Fejér-monotonicity: The proofs rely on the property of Fejér-monotonicity, demonstrating that each step of the algorithm strictly reduces the distance to the set of possible solutions.

Conclusion

This research provides a significant theoretical advancement by offering robust, self-adaptive solutions to the SCFPP for demimetric mappings. By eliminating the dependency on the operator norm and providing both weak and strong convergence guarantees, the authors have improved the state-of-the-art for inverse problems in Hilbert spaces. While the current work is purely theoretical, it lays the groundwork for more efficient multi-stage optimization and image reconstruction algorithms.

Summary updated

Abstract

The split common fixed point problem is an inverse problem that consists in finding an element in a fixed point set such that its image under a bounded linear operator belongs to another fixed-point set. In this paper, we present new iterative algorithms for solving the split common fixed point problem of demimetric mappings in Hilbert spaces. Moreover, our algorithm does not need any prior information of the operator norm. Weak and strong convergence theorems are given under some mild assumptions. The results in this paper are the extension and improvement of the recent results in the literature.
BibTeX
@article{chen2019self,
  author = {Xinhong Chen and Yanlai Song and HE Jian-ying and Liping Gong},
  title = {{Self-Adaptive Algorithms for the Split Common Fixed Point Problem of the Demimetric Mappings}},
  journal = {Journal of Applied Mathematics and Physics},
  year = {2019},
  doi = {10.4236/jamp.2019.710150},
}
Figure