Fixed Point Theory and Algorithms for Sciences and Engineering, Dec 6, 2022
In this work, we prove the weak convergence of a one-step self-adaptive algorithm to a solution o... more In this work, we prove the weak convergence of a one-step self-adaptive algorithm to a solution of the sum of two monotone operators in 2-uniformly convex and uniformly smooth real Banach spaces. We give numerical examples in infinite-dimensional spaces to compare our result with some existing algorithms. Finally, our results extend and complement several existing results in the literature.
Uploads
Papers by MARYAM ALKA
The increasing adoption of wireless sensor networks (WSNs) has brought forth significant security challenges,
particularly in mitigating distributed denial-of-service (DDoS) attacks. These attacks significantly degrade network
performance by bombarding sensor nodes with high volumes of traffic, causing resource depletion and service
interruptions. Due to the constrained processing capacity, memory, and energy of WSNs, they are highly susceptible
to such threats, which can rapidly drain battery power, reduce communication efficiency, and jeopardize data accuracy.
As a result, essential functions like real-time data collection and transmission may be disrupted, impacting critical
applications such as environmental monitoring, healthcare, and industrial automation. Moreover, DDoS attacks can
introduce security weaknesses, increasing the risk of further exploitation by malicious entities. This paper introduces
PSO_KAN, a novel hybrid model that integrates Particle Swarm Optimization (PSO) with the Kolmogorov-Arnold
Network (KAN) to improve the detection of such attacks in WSNs. Utilizing Kolmogorov’s superposition theorem;
KAN decomposes complex multivariate functions into simpler univariate components. In this framework, PSO plays a
key role in optimizing the choice and tuning of these univariate functions and their parameters to ensure accurate
representation of attack behaviors. By harnessing PSO’s ability to perform global optimization, the proposed model
delivers enhanced performance in terms of detection accuracy, scalability, and computational efficiency. The
experimental findings indicate that PSO-KAN attains a remarkable accuracy of 97.5%, surpassing conventional Neural
Networks (91.2%), Convolutional Neural Networks (94.5%), Long Short-Term Memory networks (95.1%), and even
Neural Networks optimized with PSO (96.7%). This represents a 6.3% improvement over standard Neural Networks
and a 0.8% gain compared to other PSO-enhanced models, demonstrating the superiority of the proposed method.
Additionally, the model consistently achieves high Precision, Recall, and F1-Scores across various attack types,
reinforcing its reliability in network security applications. These results underscore the effectiveness of the integration
of PSO and KAN, providing a powerful approach for strengthening the security of WSNs against sophisticated cyber
threats.
Keywords: Particle Swarm Optimization (PSO), Kolmogorov Arnold Network, Cyber Security, Distributed Denial-
of-Service, Wireless Sensor Network