Computational intelligence in industrial operations

Abstract

Computational intelligence can improve the efficiency and performance of various industrial systems. Industry 4.0 shows the rise in methods that provide (such as the Internet of Things or IoT) computational intelligence. Certainly, the rise in big data has improved the conditions for these solutions to be implemented. This research provides novel approaches to different Machine Learning (ML) and Artificial Intelligence (AI) controlled Cyber-Physical Systems (CPS) frameworks that can be implemented in a variety of industrial applications. The research illustrates the intelligent CPS framework in real-time to perform pick-and-place operations as they are one of the most widely performed operations for manufacturing. Moreover, the paper shows how latency impacts the CPS. Latency is dependent on the computational system and a networking element, which is shown by latency propagation using the sampling-resampling Bayesian modeling method. Latency provides an important basis for reproducibility in the industry. The research also provides an ML-based novel solution approach to the Job-shop Scheduling Problem (JSSP). JSSP is a combinatorial optimization problem that is NP-hard. This research provides a DL-based solution. The proposed solution is not only competitive with the Operations Research (OR)-based makespan solutions, however, it also provides generalizability which is a very important aspect for industrial applications. This research aims to optimize the supply chain network design in external dealings and provides insight into the decision-making framework. The study provides promising insight into supplier and customer negotiations that could reduce lead times and prices for customers and that customers at least should wait out the first negotiation round. Lower the auction, supplier benefits. With a lower number of auctions, suppliers can manage to sell more frequently as well.