Decision science, an interdisciplinary field that draws on disciplines including economics, statistics, computing, engineering, machine learning, psychology and management sciences, is concerned with analysing various types of information to enable better decision-making.
Industry 4.0 (I4.0) signals massive changes in the supply chain and manufacturing landscapes. Disruptive technologies for connectivity, analytics, human-machine interaction and advanced engineering have enabled a range of decision science solutions that help manufacturing and supply chain operations to “go smart”.
In this issue, PAIR discusses with Prof. George Q. HUANG, Director of the Research Institute for Advanced Manufacturing (RIAM) and Chair Professor of Smart Manufacturing, the ways in which industries benefit from technology transitions. Prof. Huang is an expert in smart manufacturing and logistics, the cyber-physical internet, and systems analytics. In May 2024, he was appointed as the new director of RIAM, leading the Institute in its pursuit of interdisciplinary research on advanced manufacturing solutions.
iFactories: The promise of smart manufacturing from hardware and software perspectives
Intelligent factories, also known as “iFactories”, utilise various technologies to improve productivity, efficiency, flexibility and decision-making in manufacturing and supply chain operations. Just as computers are composed of physical parts and computer processor chips that set operations, iFactories are built and assembled with hardware and software components. However, what gives these factories intelligence are “digital twins”, which provide virtual replications of the manufacturing processes or behaviours, so that operators can monitor the processes in real time and take corresponding actions.
“iFactories include a formal computer architecture and operating system. In the iFactory architecture, the digital twins of different manufacturing resources are used in constructing iFactories’ central processing units (CPUs) and processors. This iFactory architecture makes all physical aspects of the manufacturing process ‘visible’ and ‘traceable’, in terms of both space and time,” Prof. Huang explained. Such visibility and traceability in manufacturing is vital for operators and managers in their decisions about factory operations, planning, scheduling and execution.
The next question is how to make manufacturing “visible” and “traceable”. This necessitates the cyber-physical internet (CPI) and system analytics. CPI is a network system that allows organisation users to track materials, goods and equipment in real time, ensure timely delivery, and reduce disruptions. System analytics is a technology for analysing business procedures and processes, and identifying and suggesting improvements.
In supply chain management, small changes in demand can produce large swings in production. Such “bull-whip” effects can be addressed through CPI solutions, so that operators can avoid keeping excess inventory and using expensive warehouse space.
Achieving mass customisable production in cellular manufacturing
In manufacturing, line production is a mass production process in which the machines and workers carry out repetitive, monotonous tasks or operations in a pre-defined sequence to assemble a product that moves along the conveyor belt for further assembly until the finished product is configured. Cellular manufacturing, by contrast, involves “work stations” or “cells”, where various parts and tools are placed closer to the workers so that they can perform multiple tasks in assembling a product from start to finish. This enables the production of a wider variety of goods.
Cellular manufacturing has the advantages of boosting employee morale and supporting mass customisation, as compared to line production. However, the environment in cellular manufacturing is loaded with variables and uncertainties. Producing customisable goods means customer orders are highly diverse. Different work stations may become out of sync, as some cells may work faster than others, hence creating the problem of heterogeneous demand–capacity synchronisation (HDCS).
Prof. Huang and his team are determined to remove this bottleneck by devising solutions that help coordinate the heterogeneous capacities in the assembly cell lines in a complex manufacturing environment. The team developed a Graduation Intelligent Manufacturing System (GiMS) using artificial intelligence-enabled internet of things (AIoT) technologies and an “out of order execution” (OoOE) algorithm.
“The AIoT-enabled GiMS obtains real-time information on production status and progress, uses AI technologies to improve the dimension and granularity of the information acquired, and conducts analyses on the production capacities of individualised assembly cells using computer vision and machine learning,” Prof. Huang told us. “In this way, the heterogeneous capacity of different assembly cells can be coordinated to meet the heterogeneous customer demands effectively and efficiently, while optimising the synchronisation of production operation and management.”
The system was tested at an electronics company which develops and produces servo motors and drivers, and it significantly improved the company’s shipment punctuality and production efficiency.
More than efficiency: Putting people at the heart of the manufacturing business
The paradigm shift to smart manufacturing is about making industry not only more productive, but also more sustainable. In other words, greater value is placed on human well-being at the centre of the production process, hence making the industry more human-centric, resilient and sustainable.
Implementing ergonomics in industry is essential for creating a work environment that enhances well-being and performance. For example, work stations can be designed in a way that encourages workers to maintain good posture, thus reducing their risk of injury.
Prof. Huang and his team have developed a human-machine work system (HMWS) for proactive ergonomic risk mitigation (PERM). The smart system supports real-time data-driven human-machine synchronisation (RHYTHMS). To test the system’s performance, the team built an experimental prototype and tested it at a company that produces a wide variety of motors and motor drivers in small quantities (high-variety, low-volume production).
“The prototype integrates I4.0 enabling technologies to assist human workers in improving overall performance,” Prof. Huang explained. “The prototype includes a camera module on the top of the work station, which is used to capture and analyse operational status and processes in real time. The real-time anthropometric data is then used to construct a hyper human model, and is analysed to calculate the distances between different joints of the worker’s body. Based on the results, the racks of the machine rotate into accessible, desired positions so that the worker can easily select the needed components for assembly, hence helping to reduce workload and mitigate ergonomic risks.”
Supporting the Greater Bay Area’s development into a global hub for advanced manufacturing and modern service industries
From intelligent manufacturing system to human-machine synchronisation, the work by Prof. Huang and his team has demonstrated how technologies and complex data can be applied to bring forth smart decision science solutions that improve people’s lives and societal development.
“Informatisation has brought us new opportunities. The vast amount of data collected can make the information environment complex. But such complexity is not what we ultimately hope to achieve. Our genuine purpose should be to leverage the power of traceable, transparent real-time data in developing decision-making models that help promote the development of industries,” Prof. Huang said.
To him, the Nation’s Greater Bay Area (GBA) development stands out as a golden opportunity for research development, especially for innovations targeted at the manufacturing and logistics fields. Many practices in traditional mass manufacturing in the past may have been borrowed from, and made reference to, other countries and systems. This approach does not help achieve the Nation’s ambitious vision for GBA and smart manufacturing. Now is the prime time for researchers to make their contributions.
“The GBA is a base for informatisation. Strong manufacturing and logistics can drive developments in the modern service industries and financial industry. Researchers and developers in the field of decision science must develop related decision-making models with Chinese characteristics for the development of the GBA and industries,” Prof. Huang said.
Cyber-physical internet for synchronising cross-border logistics hubs in the Greater Bay Area
Prof. Huang is now leading a large project which aims to build a CPI called “SynchronHub” for the GBA. “CPI is an emerging system which involves ‘digitising the physical’ and ‘physicalising the digital’. We can say that CPI is a logistics system in the metaverse, or a concept which uses metaverse technology to solve logistics problems in the physical world,” he explained.
To do so, Prof. Huang and his interdisciplinary research team from PolyU and The University of Hong Kong, as well as partners from Jinan University, Shenzhen University, logistics business associations and leading companies in the GBA, are engaged in collaborative research on CPI. The developers are focusing on various components including digitisation technologies, network services and mechanisms. The resulting SynchronHub will provide decision support for synchronised logistics planning, scheduling and execution. It is hoped that with the new CPI, sending and receiving goods will become just like sending and receiving messages within chat groups using instant messaging platforms.
Leading PolyU interdisciplinary advanced manufacturing research to reach new heights
In May 2024, Prof. Huang was appointed as the new Director of the Research Institute for Advanced Manufacturing (RIAM). The Institute brings together PolyU researchers from different faculties, schools and departments for interdisciplinary research to develop advanced manufacturing solutions and industrial collaborations to put these solutions into real-world application.
“At RIAM, we shall continue our efforts and carry the RIAM torch forward to the next stage of achievements. Our vision is to establish a hub at PolyU for world-class manufacturing research and knowledge transfer to contribute to economic growth in a global context”, said Prof. Huang. “RIAM plans to support and nurture colleagues as they build and enhance critical masses for strategic research programmes with high impact, and to work together with international centres of excellence and industrial leaders to establish a global network of excellence in advanced research”.
Currently, he has initiated and is leading an RIAM project to build a large model for fashion product customisation and social manufacturing. The project combines the disciplines of mass customisation, social manufacturing, computer vision, cognitive science, large language models, and cyber-physical networking for the sake of high-performance production planning, scheduling and execution.
“Smart manufacturing requires new ways of thinking and operations to deliver intended breakthroughs. Putting together all the advanced technologies in a factory may not be necessary nor sufficient. Instead, it is imperative for multiple disciplines and key stakeholders in the supply chain to collaborate closely to innovate new processes, decision models and facilities,” Prof. Huang said.
