People

Giacomo Borghi

Address

Heriot-Watt University, Edinburgh (UK)

Email

G DOT Borghi AT hw DOT ac DOT uk

Web

TBA

Research Area

Kinetic theory for swarm-based optimization methods

Publications

  • Binary interaction methods for high dimensional global optimization and machine learning
    A. Benfenati, G. Borghi, and L. Pareschi
    Appl. Math. Optim. 86 (2022), no. 1, Paper No. 9
    doi: 10.1007/s00245-022-09836-5
    Preprint arXiv:2105.02695

  • Constrained consensus-based optimization
    G. Borghi, M. Herty, and L. Pareschi
    SIAM J. Optim. 33 (2023), no. 1, 211–236
    doi: 10.1137/22M1471304
    Preprint arXiv:2111.10571

  • A consensus-based algorithm for multi-objective optimization and its mean-field description
    G. Borghi, M. Herty, and L. Pareschi
    2022 IEEE 61st Conference on Decision and Control (CDC), Cancun, Mexico, 2022, pp. 4131-4136
    doi: 10.1109/CDC51059.2022.9993095
    Preprint arXiv:2203.16384

  • An adaptive consensus based method for multi-objective optimization with uniform Pareto front approximation
    G. Borghi, M. Herty, and L. Pareschi
    Appl Math Optim 88, 58 (2023)
    doi: 10.1007/s00245-023-10036-y
    Preprint arXiv:2208.01362

  • Repulsion dynamics for uniform Pareto front approximation in multi-objective optimization problems
    G. Borghi
    PAMM, Vol. 23, Issue 1 e202200285
    doi: 10.1002/pamm.202200285
    Preprint arXiv.2211.03378

  • Consensus based optimization with memory effects: Random selection and applications
    G. Borghi, S. Grassi, and L. Pareschi
    Chaos, Solitons & Fractals, Vol. 174, 2023, 113859
    doi: 10.1016/j.chaos.2023.113859

  • Kinetic description and convergence analysis of genetic algorithms for global optimization
    G. Borghi and L. Pareschi
    Preprint arXiv:2310.08562

  • Model predictive control strategies using consensus-based optimization
    G. Borghi and M. Herty
    Preprint arXiv:2312.13085

  • Dynamics of measure-valued agents in the space of probabilities
    G. Borghi, M. Herty, and Stavitskiy
    Preprint arXiv:2407.06389