Arman Process Journal (APJ)

Abstract The convergence of the Internet of Things (IoT) and artificial intelligence (AI), as one of the most central trends in digital transformation in the last decade, has created a new landscape in various industries, including manufacturing, transportation, energy, and smart cities. Despite the revolutionary potential of these technologies, their effective adoption and implementation face numerous challenges. The aim of this study is to provide a comprehensive framework for analyzing the factors affecting the adoption of artificial intelligence in IoT systems and to provide a proposed perspective to overcome existing barriers. This study, with an analytical-descriptive approach and using a systematic literature review, examines 46 authoritative articles published between 2018 and 2025. The theoretical framework of the study is based on the combined technology-organization-environment (TOE) model and the technology acceptance model (TAM). The literature analysis shows that the key applications of artificial intelligence in IoT include predictive maintenance, anomaly detection, real-time monitoring, and automated control. Facilitating factors include 5G infrastructure, organizational readiness, and skilled workforce, and major barriers include cybersecurity challenges, data quality, integration with legacy systems, and model scalability limitations. The proposed approach emphasizes the implementation of a hybrid architecture based on edge computing, federated learning, and software-defined networking (SDN) to overcome the existing challenges. This framework can serve as a roadmap for researchers and industry players on the path to AI-IoT-based digital transformation.

Abstract Cloud computing has emerged as a paradigm that transcends traditional distributed computing systems, such as Grid and Cluster systems, offering the capability to handle dynamic requests and diverse user requirements. As the number of users grows, there is a pressing need to deploy effective mechanisms for load balancing and task scheduling. Load balancing is essential for evenly distributing workloads across physical servers, preventing resource congestion, and enhancing overall system performance. Furthermore, considering users' service requests and the necessity for service providers to deliver accurate and timely responses, coupled with the limited resources available in the cloud environment, efficient task scheduling becomes imperative. This paper proposes an approach for optimizing virtual machine (VM) migration by combining Genetic Algorithms and Ant Colony Optimization for resource scheduling operations. Additionally, it employs K-Means clustering and fuzzy logic to quantify the dependencies between VMs and physical machines during migration, thereby maintaining load balance. The proposed model is evaluated and compared against three existing load balancing algorithms within the CloudSim simulation environment. The evaluation results demonstrate that our proposed model achieves a 4.5% reduction in task completion time, a 4.9% increase in the deadline success rate, and a 3.9% improvement in task diversity. Furthermore, computational complexity is reduced by 8.3%, VM migration efficiency is improved by 2.5%, and decision-making delay is significantly decreased by 9.5%. Additionally, the model achieves substantial energy savings of 30-35%.

Abstract The convergence of emerging technologies, including the Internet of Bio-Nano Things (IoBNT), Artificial Intelligence (AI), Digital Twins, and Nanomedicine, has opened new opportunities for the development of next-generation intelligent systems for cancer diagnosis, monitoring, and treatment. Despite significant advances in each of these fields, a comprehensive framework integrating these technologies into a unified, self-regulating, and closed-loop therapeutic ecosystem has not yet been established. This review proposes a forward-looking conceptual framework in which molecular data acquired by in vivo nanosensors are transmitted through IoBNT to a Nano-Digital Twin, analyzed using artificial intelligence models, and subsequently utilized to guide intelligent nanodrug delivery systems and therapeutic decision-making.
The principal contribution of this study is the integration of four key technologies IoBNT, Artificial Intelligence, Digital Twins, and Nanomedicine within a multilayer closed-loop therapeutic architecture that enables continuous monitoring, real-time analysis, disease progression prediction, and dynamic personalized treatment optimization. From a qualitative perspective, the proposed framework has the potential to improve diagnostic accuracy, enhance the specificity of targeted drug delivery, reduce reactive clinical decision-making, and facilitate the transition toward predictive, adaptive, and personalized medicine. From a functional perspective, the proposed architecture integrates five essential capabilities real-time monitoring, intelligent data analytics, disease progression prediction, adaptive nanodrug delivery, and continuous therapeutic feedback within a unified platform, whereas these capabilities have largely been investigated independently in previous studies.
Finally, this review discusses the major scientific, technical, and translational challenges associated with the clinical implementation of the proposed framework and presents a roadmap for the development of next-generation intelligent nanomedicine and precision medicine. The proposed conceptual framework demonstrates that the convergence of these emerging technologies has the potential to transform conventional healthcare into autonomous, intelligent, and data-driven therapeutic systems for future cancer management.

Abstract The integration of artificial intelligence into strategic decision-making processes enhances accuracy, speed, and efficiency; however, it simultaneously introduces unprecedented challenges to the traditional accountability framework within corporate governance. Employing a critical analytical approach and drawing upon findings from authoritative research in the fields of artificial intelligence and corporate governance, this article argues that existing legal and ethical systems are ill-prepared to address three structural gaps arising from algorithmic decision-making at the board level.
Accountability is defined as a tripartite relationship among an agent, the board of directors, and the general assembly of shareholders, wherein the agent is obligated to justify and explain their conduct, while the assembly holds the authority to question, judge, evaluate performance, decide on the continuation of cooperation, and take legal action if necessary. The first gap concerns the absence of a clear rule for determining responsibility when AI systems make errors, as responsibility is a "bundle" concept comprising multiple components, and merely attributing it cannot resolve ethical dilemmas. This gap manifests acutely in augmented intelligence and reinforcement learning contexts. The second gap involves the lack of a mechanism to monitor the board itself in its capacity as the overseer of AI systems—referred to herein as the "monitor of the monitor" dilemma. The third gap stems from the temporal misalignment between board decision-making cycles and AI system processing speeds, which undermines effective oversight and timely accountability.
This study proposes three policy-oriented solutions: the non-delegable rule of ultimate human responsibility; the establishment of a specialized AI committee within the board, endowed with continuous auditing authority; and the redefinition of fiduciary duty such that algorithmic literacy and oversight over the lifecycle of AI models become integral components of directors' legal responsibilities. These three solutions are analyzed within the framework of accountability's four objectives: compliance, reporting, monitoring, and enforcement.

Abstract Over the last decade, offensive cyber operations have evolved from intelligence-oriented campaigns toward destructive attacks targeting the availability and integrity of critical data and digital infrastructures. Among these threats, wiper malware has emerged as one of the most destructive cyber weapons due to its capability to permanently erase or corrupt information and disrupt organizational operations. This review paper aims to provide a comprehensive technical analysis of the design, implementation, and evolution of wiper malware by examining more than twenty documented real-world incidents, including Shamoon, NotPetya, WhisperGate, and AcidRain. Based on this analysis, a unified framework is proposed to classify data destruction techniques into three categories: physical-level destruction through storage controller commands such as ATA Secure Erase, block-level overwriting using fixed or random patterns, and logical-level corruption of metadata, partition tables, and file system structures. The study further analyzes the common architecture of modern wiper malware, including persistence mechanisms, command-and-control strategies, self-propagation techniques, parallel destruction modules, and anti-forensic capabilities. Comparative analysis demonstrates the gradual evolution of wiper malware from simple disk destruction toward sophisticated attacks targeting cloud infrastructures, virtualization platforms, and distributed storage environments. The review also identifies major technical challenges associated with forensic recovery, snapshot-based backup systems, and replicated cloud storage. As a practical contribution, the paper proposes a multilayer defensive framework based on Data Immutability, Write Once Read Many (WORM) hardware technologies, offline backup architectures, and kernel-level I/O anomaly detection to improve cyber resilience against destructive attacks. Overall, the findings highlight current technological trends, identify existing research gaps, and provide practical recommendations for developing more resilient storage architectures and effective protection mechanisms against future generations of destructive cyber threats.

Abstract The present study aims to identify and explain the key factors affecting the performance of the Golrang brand, with an emphasis on the mediating role of brand innovation, focusing on Pakshoo Company in the detergent industry. Based on the resource‑based view and the dynamic capabilities approach, market orientation, organizational learning, and internationalization were considered as organizational and strategic drivers, brand innovation as the mediating mechanism, and brand performance as the final outcome of the model.

In terms of purpose, this research is applied, and in terms of method, it is descriptive–survey and correlational. The statistical population consisted of managers and employees of Pakshoo Company located in the Alborz Industrial City of Qazvin. Out of 260 individuals, 152 respondents were selected using simple random sampling based on the Morgan table. Data were collected through a standardized questionnaire adapted from validated studies and localized for the detergent industry. The content validity of the questionnaire was confirmed by experts, and its reliability was verified using Cronbach’s alpha; the alpha coefficients for the constructs ranged from 0.712 to 0.946.

Data were analyzed using descriptive statistics, the Kolmogorov–Smirnov test, linear regression, and the Preacher and Hayes mediation model in SPSS software. The findings indicated that market orientation, organizational learning, and internationalization have a positive and significant effect on brand innovation, and brand innovation in turn positively and significantly improves brand performance. Furthermore, the mediating role of brand innovation in the relationships between the three drivers and brand performance was confirmed.

The results suggest that brand innovation acts as a connecting mechanism between organizational capabilities and developmental strategies with superior brand performance. It can serve as a lever for differentiation, market responsiveness, and strengthening competitive advantage in the detergent industry.