Arti Bandhana

Artificial intelligence is transforming various industries, including manufacturing, yet its full potential in manufacturing remains underutilized. Industry 4.0 aims to enhance productivity, operational efficiency, and decision-making, but achieving these objectives at scale remains an ongoing challenge. This paper surveys the artificial intelligence-driven integration of multi-agent systems and manufacturing execution systems as key enablers of smart manufacturing in Industry 4.0 and the emerging Industry 5.0. It reviews state-of-the-art developments, identifies key challenges, and outlines research priorities by analyzing trends from past industrial revolutions. The paper also emphasizes workforce upskilling and advocates for a problem-driven approach, prioritizing solving practical challenges over pursuing technological innovation without clear objectives. Furthermore, this paper leverages responses generated via ChatGPT and Microsoft Copilot to assess whether the discussions presented align with AI-generated insights, demonstrating an example of human–machine collaboration to set a precedent for future research. Concluding with a forward-looking agenda, it emphasizes the need for high-technology readiness level pilot implementations and stronger industry-academia collaboration to transition from theoretical breakthroughs to large-scale industrial deployment. This paper serves as a guide for researchers, policymakers, and stakeholders, providing a comprehensive perspective on technological advancements, integration challenges, and industry standards in smart manufacturing.

In peer-to-peer networks, maintaining trust is crucial. This paper introduces a novel global trust computation method using a transferable utility coalitional game, pooling local trust values. We define internal and external trust, proving our game's effectiveness in three settings compared to Eigentrust.

The ongoing rise in cyberattacks and the lack of skilled professionals in the cybersecurity domain to combat these attacks show the need for automated tools capable of detecting an attack with good performance. Attackers disguise their actions and launch attacks that consist of multiple actions, which are difficult to detect. Therefore, improving defensive tools requires their calibration against a well-trained attacker. In this work, we propose a model of an attacking agent and environment and evaluate its performance using basic Q-Learning, Naive Q-learning, and DoubleQ-Learning, all of which are variants of Q-Learning. The attacking agent is trained with the goal of exfiltrating data whereby all the hosts in the network have a non-zero detection probability. Results show that the DoubleQ-Learning agent has the best overall performance rate by successfully achieving the goal in 70% of the interactions.