当前位置: X-MOL 学术Aequat. Math. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Geometric properties of discontinuous fixed point set of ( $$\varvec{\epsilon -\delta }$$ ϵ - δ ) contractions and applications to neural networks
Aequationes Mathematicae ( IF 0.8 ) Pub Date : 2019-09-20 , DOI: 10.1007/s00010-019-00680-7
Ravindra Kishor Bisht , Nihal Özgür

In this paper, we prove some fixed point theorems under a convex combination of generalized (\(\epsilon -\delta \)) type rational contractions in which the fixed point may or may not be a point of discontinuity. As a by-product we explore some new answers to the open question posed by Rhoades (Contemp Math 72:233–245, 1988). Furthermore, we consider geometric properties of the fixed point set of a self-mapping on a metric space. We define a new kind of contractive mapping and prove that the fixed point set of this kind of contraction contains a circle (resp. a disc). Several non-trivial examples are given to illustrate our results. Apart from these, an application of discontinuous activation functions, frequently used in neural networks is also given.



中文翻译:

($$ \ varvec {\ epsilon-\ delta} $$ ϵ-δ)的不连续定点集的几何性质及其在神经网络中的应用

在本文中,我们证明了广义(\(\ epsilon-\ delta \))型有理收缩的凸组合下的一些不动点定理,其中不动点可能是也可能不是不连续点。作为副产品,我们探索了对Rhoades提出的开放性问题的一些新答案(Contemp Math 72:233–245,1988)。此外,我们考虑度量空间上自映射的不动点集的几何性质。我们定义了一种新的收缩映射,并证明了这种收缩的不动点集包含一个圆(分别是一个圆盘)。给出了几个非平凡的例子来说明我们的结果。除此之外,还给出了在神经网络中经常使用的不连续激活函数的应用。

更新日期:2019-09-20
down
wechat
bug