-
Accelerating evolutionary exploration through language model-based transfer learning
Authors:
Maximilian Reissmann,
Yuan Fang,
Andrew S. H. Ooi,
Richard D. Sandberg
Abstract:
Gene expression programming is an evolutionary optimization algorithm with the potential to generate interpretable and easily implementable equations for regression problems. Despite knowledge gained from previous optimizations being potentially available, the initial candidate solutions are typically generated randomly at the beginning and often only include features or terms based on preliminary…
▽ More
Gene expression programming is an evolutionary optimization algorithm with the potential to generate interpretable and easily implementable equations for regression problems. Despite knowledge gained from previous optimizations being potentially available, the initial candidate solutions are typically generated randomly at the beginning and often only include features or terms based on preliminary user assumptions. This random initial guess, which lacks constraints on the search space, typically results in higher computational costs in the search for an optimal solution. Meanwhile, transfer learning, a technique to reuse parts of trained models, has been successfully applied to neural networks. However, no generalized strategy for its use exists for symbolic regression in the context of evolutionary algorithms. In this work, we propose an approach for integrating transfer learning with gene expression programming applied to symbolic regression. The constructed framework integrates Natural Language Processing techniques to discern correlations and recurring patterns from equations explored during previous optimizations. This integration facilitates the transfer of acquired knowledge from similar tasks to new ones. Through empirical evaluation of the extended framework across a range of univariate problems from an open database and from the field of computational fluid dynamics, our results affirm that initial solutions derived via a transfer learning mechanism enhance the algorithm's convergence rate towards improved solutions.
△ Less
Submitted 4 February, 2025; v1 submitted 7 June, 2024;
originally announced June 2024.
-
Data-driven algebraic models of the turbulent Prandtl number for buoyancy-affected flow near a vertical surface
Authors:
Xiaowei Xu,
Andrew S. H. Ooi,
Richard D. Sandberg
Abstract:
The behaviour of the turbulent Prandtl number ($Pr_t$) for buoyancy-affected flows near a vertical surface is investigated as an extension study of {Gibson \& Leslie, \emph{Int. Comm. Heat Mass Transfer}, Vol. 11, pp. 73-84 (1984)}. By analysing the location of mean velocity maxima in a differentially heated vertical planar channel, we {identify an} {infinity anomaly} for the eddy viscosity $ν_t$…
▽ More
The behaviour of the turbulent Prandtl number ($Pr_t$) for buoyancy-affected flows near a vertical surface is investigated as an extension study of {Gibson \& Leslie, \emph{Int. Comm. Heat Mass Transfer}, Vol. 11, pp. 73-84 (1984)}. By analysing the location of mean velocity maxima in a differentially heated vertical planar channel, we {identify an} {infinity anomaly} for the eddy viscosity $ν_t$ and the turbulent Prandtl number $Pr_t$, as both terms are divided by the mean velocity gradient according to the standard definition, in vertical buoyant flow. To predict the quantities of interest, e.g. the Nusselt number, a machine learning framework via symbolic regression is used with various cost functions, e.g. the mean velocity gradient, with the aid of the latest direct numerical simulation (DNS) dataset for vertical natural and mixed convection. The study has yielded two key outcomes: $(i)$ the new machine learnt algebraic models, as the reciprocal of $Pr_t$, successfully handle the infinity issue for both vertical natural and mixed convection; and $(ii)$ the proposed models with embedded coordinate frame invariance can be conveniently implemented in the Reynolds-averaged scalar equation and are proven to be robust and accurate in the current parameter space, where the Rayleigh number spans from $10^5$ to $10^9 $ for vertical natural convection and the bulk Richardson number $Ri_b $ is in the range of $ 0$ and $ 0.1$ for vertical mixed convection.
△ Less
Submitted 5 April, 2021;
originally announced April 2021.
-
Influences of initial streamwise rotation of a droplet under a uniform flow
Authors:
Eric K. W. Poon,
Shaoping Quan,
Jing Lou,
Andrew S. H. Ooi
Abstract:
A spherical droplet is given an initial rotation in the streamwise direction and is impulsively accelerated by a uniform free stream. Numerical results for the deformation and dynamics of the droplet are obtained by utilising a finite volume staggered mesh method with a moving mesh interface tracking scheme. The fluid dynamics videos of the droplet are presented in the Gallery of Fluid Motion, 201…
▽ More
A spherical droplet is given an initial rotation in the streamwise direction and is impulsively accelerated by a uniform free stream. Numerical results for the deformation and dynamics of the droplet are obtained by utilising a finite volume staggered mesh method with a moving mesh interface tracking scheme. The fluid dynamics videos of the droplet are presented in the Gallery of Fluid Motion, 2010. By initially rotating the droplet in the streamwise direction, the droplet deforms differently depending on the non-dimensional rotation rate, $Ω^*$. The families of droplet shape are, in ascending $Ω^*$, biconvex, convex-concave and biconcave. While the biconvex and convex-concave families are due to the compression by the combined vortex ring (the lee side of the interface). The biconcave family is formed because of the huge surface tension at the droplet edge that restrains further deformation there. In a special case, the biconcave family releases the combined vortex ring from the droplet. The release of the vortex ring is observed to be at $Re$ one order of magnitude lower than a similar mechanism observed for the flow past a solid sphere. This release of the vortex ring has led to a series of events and a reduction in $C_D$.
△ Less
Submitted 20 October, 2010; v1 submitted 15 October, 2010;
originally announced October 2010.
