Research Projects

Introduction: With the ever changing marketplace, globalization, and the increasing impact of manufacturing on our environment, there is a paramount need to refine our existing manufacturing ecosystem, as well as develop newer paradigms. The evolution and adaption of Industry 4.0 – the digital revolution in the manufacturing field – is a testament to the commitment of the global manufacturing community to better serve the costumer and reduce the adverse impacts on our environment. Central idea: This project leverages on the core idea (data-centric decisions) within Industry 4.0 to develop an efficient manufacturing ecosystem that is decentralized, digitally connected, agile and resilient, with capability to collectively fabricate multiple types of products. More precisely, we propose to design data-driven algorithms and control-decision systems that will assist spatially separated, fabrication facilities with variety of capabilities and capacities to work as a single manufacturing unit - a cyber-physical manufacturing system - that could be spread over an entire state or country. The proposed cloud based manufacturing system will onboard the dedicated manufacturing units (DMUs) and the raw material units (RMUs) onto a single digital platform, with comprehensive details on the capabilities of each units. As and when, a customer orders a product, the process flow will start from an RMU and continue through a (parallel) sequence of DMUs, with more RMUs joining the sequence at separate stages by providing the requisite raw material, before the finished product finally reaches the customer. Such a setup serves multifold benefits such as customization, regional jobs, lower transportation costs, reduced lead times, resilience to unforeseen disasters (as big as the COVID-19 pandemic), and many others. Challenges and proposed solution: A critical key to the success of such a decentralized ecosystem is an efficient cyber layer, which involves collecting appropriate data at multiple points of the manufacturing network and developing robust control and decision systems based on the available data. In particular, to harness the capability of the decentralized RMUs and DMUs, and enable them to work as a one large production unit, it is crucial to develop data-driven methodologies to solve the underlying planning, scheduling, routing, resource allocation, and assignment problems. Each of these problems fall into the category of combinatorial optimization problems (optimization over discrete space), and even individually, are difficult to solve (many of which are NP-hard). In this project we propose to develop efficient cyber layer for a decentralized manufacturing setup that takes data-driven decisions to address the underlying coupled combinatorial optimization problems. In particular, we'll develop a digital platform and data-based learning systems for the cyberphysical manufacturing systems.