This research explores how Large Language Models can fully automate the generation of combinatorial optimization systems, reducing the need for expensive expertise in building decision support tools.

• Addresses the historical challenge of limited reusability in optimization systems
• Makes decision support more accessible for small and medium enterprises
• Eliminates the need for bespoke software development for each unique optimization problem
• Particularly valuable for engineering applications where optimization can improve efficiency and reduce costs

This advancement matters for engineering teams by democratizing access to powerful optimization tools that were previously available only to organizations with specialized expertise and resources.

Fully Automated Generation of Combinatorial Optimisation Systems Using Large Language Models