BOSS-2026
Bayesian and surrOgate-aSsisted Search and Optimisation - PPSN workshop 2026
In many real-world optimisation problems, evaluating the objective function(s) can be costly, in terms of time and/or money. Surrogate-assisted optimisation attempts to alleviate this problem by utilising comparatively cheap surrogate models to estimate objective function value(s) at any location in decision space and, potentially, capture any uncertainty in those predictions. Techniques based on Bayesian Optimisation (Efficient Global Optimisation), and surrogate-assisted approaches more generally, have been employed in tackling a wide variety of optimisation problems. These include both single- and multi-objective optimisation problems, as well as those with time-varying objectives, constraints, solution robustness requirements, and problems with both deterministic and stochastic outputs.
There are many great success stories of using surrogate-based optimisation strategies to solve some of the most expensive real-world problems, such as hyperparameter tuning of machine learning models, neural architecture search, material design, reinforcement learning, design optimisation in computational fluid dynamics, and drug discovery. Despite the numerous successful applications of surrogate-based optimisation methods, there are still fundamentally challenging aspects to both applying existing methods to new problems and developing new methods to overcome outstanding methodological hurdles. These include effectively modelling objective functions with non-smooth characteristics, spatial-dependent length-scales, and high-dimensional input spaces, designing effective and efficient acquisition functions, utilising the gradient and the Hessian matrix, as well as taking into account the preferences of the decision maker during the optimisation, i.e., bringing the human back into the optimisation loop.
The workshop will focus on the sharing of best practice, examining the latest methodological improvements and discussing Bayesian Optimisation and other surrogate-assisted optimisation strategies, particularly those that take inspiration from nature. In doing so, the organisers hope to foster new collaborations among researchers and practitioners, bridging the gaps between theoretical advances and practical applications. The Workshop on Bayesian and surrOgate aSsisted Search and Optimisation (BOSS) to be held at PPSN 2026 in Trento, Italy, aims to promote the research on Bayesian Optimisation and surrogate-assisted optimisation, particularly those that take inspiration from nature. Topics of interest include (but are not limited to):
Topics of interest
- Bayesian optimisation.
- Advanced machine learning techniques for constructing surrogates
- Model management in surrogate-assisted optimisation
- Multi-level, multi-fidelity surrogates
- Complexity and efficiency of surrogate-assisted methods
- Small and big data-driven evolutionary optimisation
- Model approximation in dynamic, robust and multi-modal optimisation
- Model approximation in multi- and many-objective optimisation
- Surrogate-assisted evolutionary optimisation of high-dimensional problems
- Comparison of different modelling methods in surrogate construction
- Surrogate-assisted identification of the feasible region
- Comparison of evolutionary and non-evolutionary approaches with surrogate models
- Test problems for surrogate-assisted evolutionary optimisation
- Performance improvement techniques in surrogate-assisted evolutionary computation
- Performance assessment of surrogate-assisted evolutionary algorithms
Contact
If you have any enquiries or suggestions about BOSS, please email contact: Paul Kent