Event Details

CENTRAL Seminar

10:00 – Vinh Doan (University of Warwick): “Stochastic congestion games under random participation”

10:50 –11:10 – Coffee break

11:10 – Ahmed Kheiri (University of Manchester): “Hyper-heuristic for Complex Logistics Operations”

12:00 – Conclusion

Abstract: Vinh Doan (University of Warwick)

We study stochastic weighted congestion games, in which players with heterogeneous weights participate independently with known probabilities, capturing demand uncertainty in systems such as cloud computing and global shipping logistics. We analyse the equilibria and efficiency via price of anarchy (PoA) of the games under two cost sharing

rules, the Shapley value (SV) and proportional sharing (PS) rule, for polynomial cost functions. The results demonstrate how random participation can mitigate worst-case equilibrium inefficiency in weighted congestion games. This is joint work with Weicheng Wang and Bo Chen.

Abstract: Ahmed Kheiri (University of Manchester)

Real-world routing and scheduling problems are often characterised by complex operational constraints, dynamic decision environments and competing business objectives that make exact optimisation impractical. Hyper-heuristics offer a higher-level search methodology that aims to automate the selection and coordination of problem-specific heuristics, reducing the need for extensive manual algorithm design while maintaining strong solution quality. This talk presents the application of selection hyper-heuristics to industrial routing and scheduling problems, with a focus on large-scale inventory routing and logistics operations. Using a real-world case study from the industrial gas distribution sector, I will demonstrate how sequence-based hyper-heuristics can learn effective combinations of low-level heuristics during the search process and adapt their behaviour according to observed performance.