AI-OPT 2025 is an informal yet highly specialised forum aimed at bringing together researchers across a wide spectrum of the AI-driven optimisation domain. The event created a collaborative environment for exchanging ideas, discussing cutting-edge research, and fostering new professional relationships in the optimisation community. My participation in this workshop was motivated by two primary objectives: to present my recent research on evolutionary bi-objective optimisation for the stochastic multiple knapsack problem in dynamic environments, and to connect with peers and senior researchers whose expertise intersects with my own work.

Participation in AI-OPT 2025 was directly aligned with my PhD research focus, which lies at the intersection of evolutionary computation, metaheuristics, and real-world optimisation problems. My work aims to address the challenges of solving complex combinatorial optimisation problems in the presence of dynamic constraints and stochastic variability.

The two-day program was designed to balance technical depth with opportunities for discussion and collaboration. The agenda included a series of technical presentations, invited talks from leading researchers, and early-career researchers. This balance between formal and informal engagement was one of the most valuable aspects of the event.

I presented my research entitled “Bi-objective optimisation for the stochastic multiple knapsack problem in dynamic environments”. The presentation focused on the development and evaluation of multi-objective evolutionary algorithms for addressing uncertainty in combinatorial optimisation problems. Specifically, I discussed algorithmic approaches that integrate stochastic modelling with bi-objective problem formulations to optimise both expected performance and robustness under changing conditions. The talk also addressed practical challenges in handling dynamic constraints and balancing computational efficiency with solution quality.

The audience provided constructive and insightful feedback, which has given me new perspectives for refining my approach. Several attendees offered suggestions on alternative modelling frameworks, as well as ideas for benchmarking the proposed methods against more diverse problem scenarios. This feedback will be very helpful in shaping the next stages of my research, particularly in improving the adaptability of my algorithms to different classes of real-world optimisation problems.

Many of the presentations at AI-OPT 2025 were directly relevant to my work. For example, talks on adaptive operator selection in evolutionary algorithms, multi-objective optimization, and large-scale combinatorial problem solving offered both theoretical insights and practical techniques that I can integrate into my own models. Sessions exploring real-world applications—from logistics and energy optimisation to scheduling in dynamic environments—were particularly inspiring, as they highlighted the versatility of AI-driven optimisation methods beyond my current domain. These discussions sparked several ideas for extending my research into new problem domains and for developing hybrid algorithmic frameworks that combine my existing approaches with complementary techniques.

Beyond the technical program, the workshop offered numerous opportunities for networking. Informal discussions during breaks and lunches provide opportunities for building relationships with fellow researchers. I was able to connect with both early-career and senior academics.These conversations also provided a broader perspective on career pathways in academia and industry, as well as funding opportunities for joint research initiatives.

The impact of attending AI-OPT 2025 has been both immediate and long-term. In the short term, the technical feedback and exposure to new methodologies will directly inform my ongoing experiments and modelling choices. In the longer term, the connections I have made are likely to lead to collaborative projects, joint publications, and knowledge exchange that will strengthen my research outcomes. Additionally, I plan to share the key insights gained at the workshop with my research group, ensuring that the benefits of my participation extend beyond my individual work.

Attending AI-OPT 2025 has also reinforced my confidence in the relevance and potential impact of my research. The strategies and ideas I encountered will be directly applied to improve my optimisation models, with a particular focus on enhancing their scalability, adaptability, and robustness.

In conclusion, AI-OPT 2025 was an invaluable experience that significantly advanced my research direction, expanded my professional network, and provided actionable strategies for enhancing my optimisation approaches. I am sincerely grateful to OPTIMA and the organization committee for the financial and organisational support that enabled my attendance. The outcomes of this travel will directly contribute to the advancement of my PhD research, strengthen collaborative connections within the AI-based optimisation community, and support my ongoing development as a researcher in this dynamic and impactful field.