Pacific Optimisation Conference 2023: A Fusion of Ideas and Innovation
The Pacific Optimisation Conference 2023 recently concluded. It was a resounding success, uniting luminaries, researchers, and practitioners at the forefront of optimisation and artificial intelligence (AI). Themed "Optimisation and AI in Real-World Applications," the conference showcased insightful plenary sessions alongside a diverse array of parallel sessions, providing a comprehensive exploration of cutting-edge developments.
Intelligent Control for Low-Carbon Operation of Energy Intensive Equipment
The conference kicked off with Professor Dr Tianyou Chai (Northeastern University), who gave an in-depth exploration of the challenges faced by China's extensive process industry. The first plenary session introduced groundbreaking research ideas utilising Cyber-Physical Systems (CPS) and Industrial Artificial Intelligence (AI). Attendees gained insights into an intelligent control system designed to achieve low-carbon operation by combining mechanism analysis with deep learning. The system was successfully applied to a fused magnesium furnace and remarkably reduced carbon emissions.
Entirely Actuated System Approach for Control: Unleashing the Power of FAS
The Fully Actuated System (FAS) approach took centre stage in the second plenary session, showcasing its optimal control capabilities for complex nonlinear dynamical systems. Professor Guangren Duan (Southern University of Science and Technology; Harbin Institute of Technology) highlighted the explicit solutions for control vectors provided by FAS, making it particularly effective for handling time-varying nonlinear systems and nonholonomic constraints. Attendees gained a new perspective on system analysis and design, acknowledging the power of FAS in solving optimal control issues.
Data-driven Decision Support in Complex Situations—IRIS Integrated Reachback Information System: Case Study OPERA: Operational Analysis and Sector-based Optimization with IDEA4C
Amidst the intricate challenges decision-makers face in politics and the logistics industry, the third plenary session by Professor Stefan Wolfgang Pickl (Bundeswehr University Munich) addressed the pressing need for optimised supply chains amidst disruptions. The session presented the "OPERA - Future Operations" project, unveiling an innovative interactive visualisation tool designed to assist political decision-makers in exploring and adapting to diverse strategies. Additionally, the talk introduced the IDEA4C approach within the IRIS project, outlining a sector-based methodology for identification, analysis, and adaptation, forming a robust integrated modelling suite for relief distribution strategy development. This 4C approach is a groundbreaking hybrid optimisation framework developed collaboratively with disaster and emergency management experts to tackle real-world crisis challenges.
Nonsmooth Analysis and Sparse Optimization: Unlocking Efficiency in Large-Scale Problems
The fourth plenary session by Professor Defeng Sun (The Hong Kong Polytechnic University) unveiled the critical role of nonsmooth analysis in addressing large-scale sparse optimisation problems. Fundamental concepts such as local Lipschitz functions and the Moreau-Yosida regularization were introduced, showcasing the efficiency of nonsmooth systems in solving constrained optimisation problems. Practical applications, including machine learning models like convex clustering and lasso, were discussed, contributing to ongoing advancements in software packages for large-scale conic optimisation problems.
Process Flexibility and Supply Chain Risk Mitigation: Navigating Uncertainties
Against international tensions and pandemic disruptions, the fifth plenary session by Associate Professor Dr Chou Cheng-Feng (National University of Singapore) explored the evolving landscape of global supply chains. The session emphasised the effectiveness of process flexibility in mitigating supply chain uncertainties. Methodologies for analysing supply network resilience were discussed, providing insights into designing more resilient supply chains by leveraging process flexibility to adapt to changing market conditions.
Recent Trends in Descent Methods for Multiobjective Optimization: Convergence in Complexity
The concluding plenary session by Associate Professor Dr Ellen Hidemi Fukuda (Kyoto University) delved into the evolution of descent methods for multiobjective optimisation problems since the turn of the century. Attendees were guided through the basics of descent methods, including subproblem considerations and measurement of Pareto efficiency. Recent developments, especially in proximal gradient methods and their accelerated versions, offered a glimpse into the future of optimisation, focusing on composite structures.
Parallel Sessions: A Tapestry of Knowledge
Complementing the plenary sessions, the conference featured a rich tapestry of parallel sessions covering a spectrum of topics:
• Optimisation Methods
• Control Theory
• Optimal Control Theory
• Optimisation Methods
• Applications in Engineering and Finance
• Supply Chain Optimisation, Graph Theory, and Number Theory
• Applications in Engineering and Finance, Artificial Intelligence, Statistics, and Machine Learning
These parallel sessions facilitated focused discussions, allowing attendees to tailor their conference experience to their interests and expertise. The interactive nature of the parallel sessions fostered interdisciplinary collaboration, allowing participants to engage deeply with the diverse facets of optimisation and AI.
The Pacific Optimisation Conference 2023 was a melting pot of ideas and innovation, fostering collaboration and dialogue at the intersection of optimisation and AI. The plenary and parallel sessions collectively contributed to the continuous evolution of optimisation and AI technologies, leaving a lasting imprint on the landscape of real-world applications.
Associate Professor Dr Jason Ng Wei Jian
School of Mathematical Sciences
Email: @email