Optimisation and Decision Systems

employees

publications

projects

Optimization and decision-making techniques are essential for navigating and managing the multiple challenges and complexities of today’s world. These methodologies underpin the development of optimized solutions required by various industries and support informed decisions in dynamic, unpredictable environments. Their applications span structural mechanics, logistics, business processes, environmental management, telecommunications, new materials design, healthcare, and many more.

Optimization techniques

It empowers partners to enhance the competitiveness and performance of their processes, products and/or services. They can significantly reduce development and production costs, shorten the required time-to-market, and accelerate innovation by facilitating the discovery of novel and creative solutions. Optimization techniques allow for an efficient exploration of design spaces, consideration of complex trade-offs and identification of optimal solutions, even for the most complex and constrained problems. These techniques can be integrated with Smart Design of Experiment approaches, sensitivity analysis and surrogate models for a comprehensive and intelligent exploration of the design space. When combined with uncertainty quantification, optimization can be further enhanced to ensure robust and reliable products.

Decision support systems

They play a pivotal role in aiding partners in their everyday decision-making processes. These systems provide a structured method to rank alternative solutions generated by the optimization process in a coherent and judicious manner. Well-defined decision support systems can simultaneously consider all conflicting goals, avoiding the pitfalls of widely used but inherently flawed weighted functions or similar approaches.

These techniques create both opportunities and challenges. Despite their clear benefits, many sectors still apply them partially or inadequately. This lack of solid foundational understanding limits their effective application and broader adoption in real-world scenarios. Addressing these challenges requires a concerted effort to build and disseminate robust foundational knowledge, ensuring that optimization and decision-making techniques are applied effectively and to their full potential.

Mission

The Optimization and Decision Systems group is dedicated to making optimization and decision-making widely competitive and accessible in practical applications, with the following objectives:

Support organizations in significantly improving their decision-making processes through the use of state-of-the-art multi-objective optimization methods and tools.

Assist engineers and technical professionals in adopting these advanced technologies, democratizing access to information and software tools, and fostering interaction between professionals from different departments.

Contribute to the development of sustainable products, services and processes by optimizing current business practices.

Collaborate with researchers across various fields (health, logistics, defence, manufacturing and others) to drive meaningful social changes by supporting well-designed products using optimization and decision-making techniques.

Aid investors in achieving improved returns on investment by optimizing products, services and processes. Apply optimization and decision-making tools to develop lean processes, more efficient and sustainable products and services with reduced production costs and lower energy consumption.

Enhance the value of computer infrastructure by employing surrogate models and digital twins to reduce the costs associated with model evaluations.

Reduce risks by quantifying uncertainties related to specific manufacturing processes and environmental conditions, ensuring the design of robust and reliable products and services.

Support organizations in using multi-disciplinary optimization techniques to identify the Pareto front with optimal environmental and socio-economic impacts, contributing to the sustainable development of products, services and processes.

Minimize overall environmental impact and life cycle costs while improving sustainability performance by integrating life cycle assessment (LCA) and associated uncertainties into the optimization and decision-making tools from the early stages of product and service development. Emphasize circularity in manufacturing and the effects within the digital thread.

Scope of Expertise

The group possesses extensive expertise in various areas related to optimization and decision-making. This includes both the mathematical and computational aspects of optimization, particularly combinatorial optimization and decision-making, with strong interactions in machine learning and simulation.

In the area of combinatorial optimization, the group has expertise in modelling real-world problems using integer linear programming techniques and in developing efficient exact and heuristic algorithms to solve them. The group is also experienced in constrained and global optimization, multi-agent simulation and industrial applications in logistics, manufacturing and supply chain management. Recently, its expertise has expanded to include health, business processes and integration with digital twins.

The group is exploring new approaches to address uncertainty in complex engineering design problems. By combining state-of-the-art optimization techniques, it can efficiently tackle large-scale problems involving multiple objectives, disciplines, mixed discrete and continuous variables, and uncertain quantities, all aimed at supporting well-informed decision-making processes.

Research activities also focus on developing and improving techniques to reduce the computational cost of deriving accurate uncertainty quantifications in large multi-objective and multi-disciplinary optimization problems. This involves the use of surrogate models, digital twins and multi-fidelity optimization techniques.

In more detail, the group works on:

Evaluating methods and tools to effectively handle optimization with numerous objectives and constraints in continuous domains, incorporating the proper design of experiments and sensitivity analysis steps.

Addressing the challenges of formulating problems that involve multiple disciplines, identifying interdisciplinary couplings and managing uncertainty.

Tackling multi-scale supply chain problems requiring simultaneous spatial and temporal optimization, considering sustainability, uncertainty across temporal scales and large-scale complexity. Evaluating new methods and tools and investigating new design criteria for real-world applications to build optimized, global and interconnected logistics systems.

Developing methods and tools to manage the interaction between uncertainty quantification and optimization, finding the ideal balance between enhancing reliability and reducing resource utilization.

Leveraging large datasets for optimization and decision learning, handling varying data quality levels, and evaluating machine learning methods and tools to process big data.

Integrating sustainability, energy efficiency and resource consumption into optimization processes, aiming to maximize economic potential, minimize environmental impact and enhance social benefits, as well as incorporating life cycle assessment methodologies into design problems.

Investigating the role of digital twins in scaling up metamodelling for optimization and decision-making.

Addressing optimization challenges in manufacturing, including additive manufacturing and new materials, integration with current CAD/CAE systems, and promoting circular manufacturing and digital thread applications.

Selecting and defining the best surrogate models for optimization problems, utilizing sensitivity analysis methods to develop precise and efficient surrogate models, and addressing non-continuous problems.

Developing strategies to tackle complex or intractable problems involving binary variables and non-convex spaces in terms of objectives.

Collaborating with environmental research groups to apply optimization techniques to new environmental regulations and challenges.

Utilizing numerical optimization to design new materials with complex and conflicting performance metrics, addressing complex constraints.

Exploring distributed decision-making processes that do not rely on centralized authority, opening new research options to investigate them.

Combining decision-making notation (DMN) with business process notation (BPMN) to represent and optimize business processes.

Evaluating and validating methods to rank alternative solutions produced by optimization processes, supporting decision-making by considering all conflicting goals simultaneously.

Applying multi-objective optimization in the health domain, particularly in decision support and improving patient conditions, leveraging data from sensor devices.

Integrating optimization with AI and ML across various domains, addressing challenges from algorithm complexity to ethical considerations.