ABSTRACT: 

This paper presents a Hybrid Edge–Cloud Embedded AI Architecture designed to enhance the performance of smart communication systems by addressing key challenges associated with latency, energy consumption, and scalability. The proposed framework integrates three hierarchical layers: device, edge, and cloud, enabling distributed intelligence and efficient task execution across the network. At the device layer, embedded AI (TinyML) enables local data pre-processing and inference, reducing unnecessary data transmission. The edge layer provides real-time processing, low-latency decision-making, and adaptive task offloading, while the cloud layer supports large-scale data analytics, global system coordination, and AI model training. A multi-objective optimization model is developed to intelligently allocate computational tasks among the layers based on latency, energy efficiency, and computational cost. Simulation results demonstrate that the hybrid architecture significantly reduces system latency and energy consumption compared to traditional cloud-centric approaches. The results further show that the system maintains stable performance under increasing workloads, confirming its scalability and adaptability to dynamic network conditions. The edge layer is identified as a critical component for balancing real-time responsiveness and computational efficiency. Overall, the proposed architecture provides a robust, scalable, and efficient solution for next-generation communication systems, with potential applications in Internet of Things (IoT), smart cities, industrial automation, and autonomous systems.

Keywords: Hybrid Edge–Cloud Computing, Embedded AI, Task Offloading, Smart Communication Systems, Edge Intelligence, Latency Optimization, and Energy Efficiency.

ABSTRACT: 

The increasing complexity of cloud-native and distributed computing environments has made traditional fault management approaches insufficient for maintaining reliability, availability, and service continuity. Modern cloud infrastructures comprise large-scale, heterogeneous, and dynamically changing resources that require intelligent mechanisms capable of autonomous adaptation and recovery. This paper presents an Autonomic Distributed System (ADS) for designing self-healing cloud architectures based on the Monitor–Analyze–Plan–Execute over a Knowledge Base (MAPE-K) framework enhanced with artificial intelligence and distributed orchestration. The proposed architecture integrates real-time monitoring, machine learning-based anomaly detection, intelligent fault diagnosis, adaptive recovery planning, and automated execution mechanisms within a closed-loop control framework. A reinforcement learning-inspired decision model is incorporated to support autonomous resource management and recovery optimization under dynamic operating conditions. The architecture is evaluated through a simulation-based framework using key performance indicators including system latency, failure behaviour, recovery time, availability, and service-level agreement (SLA) compliance. Results demonstrate that the proposed self-healing architecture effectively detects and isolates faults, maintains bounded latency under workload fluctuations, achieves rapid and consistent recovery following failures, minimizes SLA violations, and prevents fault propagation across distributed components. The findings confirm that integrating autonomic control principles with AI-driven intelligence significantly enhances cloud resilience, operational efficiency, and service reliability. The proposed framework provides a scalable foundation for next-generation self-managing cloud and edge-cloud systems capable of sustaining dependable operation in highly dynamic and mission-critical environments.

Keywords: Autonomic Computing; Self-Healing Systems; Distributed Systems; Cloud Computing; MAPE-K; Artificial Intelligence; Machine Learning.

ABSTRACT: 

Sound is considered as a fundamental element of the architecture of a modern museum. The article analyzes how the transition to Audio over IP protocols transforms the design of multimedia complexes. Based on three typical scenarios (a stationary atrium, a temporary exhibition and a mobile module), it is demonstrated how a single network infrastructure ensures the stability, flexibility and scalability of the sound environment. It is shown that the Dante and AES67 protocols make it possible to integrate historical musical instruments, object‑oriented acoustic installations and mobile multimedia platforms into the overall digital contour of the building. As a result, the museum audio environment is presented as a single digital system capable of adapting to curatorial tasks, operational constraints, and various forms of communication.

Keywords:  Audio over IP, Dante, AES67, museum acoustics, multimedia expositions, network protocols, digital infrastructure, multimedia complex.

Keywords:   Analytic Hierarchy Process; Fuzzy Logic; Workforce Allocation; Multi-Criteria Decision-Making; Triangular Fuzzy Numbers; Jordan Industrial Sector.

ABSTRACT: 

Effective workforce allocation is central to industrial productivity, yet the decision-making tools most used to support it struggle to accommodate the uncertainty and subjectivity inherent in expert judgment. This study develops and validates a Fuzzy Analytic Hierarchy Process (Fuzzy-AHP) model for workforce allocation in the Jordanian industrial sector, a context shaped by rapid Industry 4.0 transformation, persistent youth unemployment, and a widening gap between the skills industry demands and the skills the labor market supplies. The problem addressed is that the classical Analytic Hierarchy Process (AHP) requires decision-makers to express pairwise preferences as exact numerical values, a requirement that does not reflect the imprecise nature of real managerial judgment. To resolve this limitation, the study integrates fuzzy logic, represented through triangular fuzzy numbers, into the classical AHP framework using Buckley's geometric mean method. Data were collected through semi-structured interviews and pairwise-comparison surveys with six experts’ production managers, plant managers, a CEO, and operations engineers representing pharmaceutical, construction-products, and cement companies operating in Jordan. Thematic analysis of the interviews, cross-checked against the literature, identified six criteria: Experience, Technical, Human and Behavioral, Operational, Organizational, and Logistical factors. All six individual pairwise matrices achieved Consistency Ratios between 0.0142 and 0.0229, confirming the reliability of expert judgments. Geometric-mean aggregation produced crisp AHP weights led by Experience (37.7%) and Operational factors (21.7%), while the Fuzzy-AHP extension, verified through alpha-cut sensitivity analysis at five confidence levels, confirmed the same ranking and produced defuzzified weights of Experience (36.9%), Operational (21.8%), Technical (18.5%), Human and Behavioral (11.2%), Organizational (6.6%), and Logistical (5.0%). The convergence between the crisp and fuzzy results validates the robustness of the derived priorities and demonstrates that the Fuzzy-AHP model provides Jordanian industrial firms with a transparent, uncertainty-aware, and practically actionable tool for workforce allocation decisions.

Keywords:   Analytic Hierarchy Process; Fuzzy Logic; Workforce Allocation; Multi-Criteria Decision-Making; Triangular Fuzzy Numbers; Jordan Industrial Sector.

ABSTRACT: 

Differential equations have long provided a fundamental mathematical framework for describing the evolution of physical, biological, and engineering systems. Nevertheless, many real-world processes exhibit memory, nonlocal interactions, evolving dynamic characteristics, uncertainty, and data-dependent behavior that cannot be adequately represented by conventional models with fixed and precisely known parameters. Fractional differential equations introduced an important extension by providing mathematical mechanisms for representing memory and nonlocality, while variable-order fractional models further enabled the representation of systems whose memory characteristics evolve over time or space. In parallel, fuzzy differential modeling provided mechanisms for representing imprecision and uncertainty, leading to the development of fuzzy fractional differential equations and fuzzy-order fractional calculus, where uncertainty can extend to the fractional order itself. More recently, data-driven and neural approaches have begun to infer or learn variable fractional orders directly from observations. This paper develops a research roadmap that organizes these developments into a sequential modeling hierarchy extending from classical differential equations to autonomous memory-aware dynamical systems. The proposed progression is not intended as a strict historical chronology or as a claim that each level has completely replaced the previous one. Instead, each level represents an additional modeling capability: dynamic change, memory, evolving memory, uncertainty, uncertain memory, uncertain memory characteristics, data-driven identification, neural adaptation, learnable uncertain memory, adaptive memory modeling, and ultimately autonomous memory-aware dynamics. The study synthesizes established developments in fractional calculus, variable-order modeling, fuzzy differential equations, fuzzy-order formulations, and learning-based fractional dynamics, and identifies research gaps at the interfaces between these areas. Particular attention is given to the transition from prescribing memory characteristics to identifying, learning, quantifying, and adapting them from data. The roadmap reveals a broad research space for applied mathematics and engineering, including intelligent control, structural health monitoring, energy systems, biomedical engineering, materials, transportation, robotics, and complex network dynamics. The paper proposes that the next generation of differential models should increasingly move from models that merely contain memory toward models capable of discovering and adapting their representation of memory.

 Keywords:  Differential equations; Fractional differential equations; Fractional calculus; Memory effects; Variable-order fractional dynamics; Fuzzy differential equations; Fuzzy-order fractional calculus; Data-driven identification; Neural fractional dynamics; Adaptive memory; Intelligent dynamical systems; Engineering applications.