Deep Learning-Enabled Fault-Tolerant Task Scheduling and Security Framework for Maritime Engineering Cloud Systems

Abstract

Maritime engineering is rapidly transforming with the integration of intelligent computational technologies that enhance vessel operations, reduce operational costs, and improve overall system performance. However, unforeseen failures during mission-critical maritime processes—such as navigation, cargo handling, and fleet coordination—can significantly degrade system reliability. To address these challenges, this research proposes a novel Dynamic Clustering with Flamingo Searchbased fault-tolerant scheduling mechanism (DCFS-FTDSS) designed specifically for maritime applications. The model improves operational scheduling in maritime cloud systems by maximizing resource efficiency and enabling real-time fault tolerance. In addition, a Queuing Theory-driven Deep Neural Network (QT-DNN) is introduced to manage dynamic resource allocation across maritime communication and control systems, while considering parameters such as bandwidth fluctuations, vessel workload, and communication capacity. Binary Bird Swarm Optimization (BBSO) is incorporated to further enhance decision-making, ensuring high Quality of Service (QoS) for maritime networks. To mitigate disruptions caused by malicious onboard or external nodes—such as sensor spoofing, vessel cybersecurity breaches, or communication tampering—the model integrates a Poaching Raptor Optimization-based unsupervised detection method. This security framework employs a dual-module approach consisting of data trust evaluation and reputation scoring, ensuring reliable information exchange and minimizing false detection of compromised nodes.All simulations and performance evaluations were carried out using MATLAB. The experimental findings demonstrate a significant improvement in resource allocation accuracy, achieving an efficiency of 98.1%. Additionally, the methodology achieved notable energy efficiency, reducing consumption to 0.506, a critical factor for sustainable and cost-effective maritime operations. The proposed framework introduces an advanced and resilient architecture for maritime computing systems, offering valuable contributions toward enhancing operational reliability, fault tolerance, and autonomous decision-making in next-generation marine technologies.