An interacting particle consensus method for constrained global optimization
This addresses optimization challenges in engineering and scientific domains where traditional gradient-based methods fail, though it appears incremental as an extension of consensus-based optimization.
The paper tackles constrained global optimization problems with non-differentiable or non-convex loss functions by developing a particle-based method that incorporates a forcing term for constraints, establishing theoretical convergence and demonstrating performance through numerical experiments.
This paper presents a particle-based optimization method designed for addressing minimization problems with equality constraints, particularly in cases where the loss function exhibits non-differentiability or non-convexity. The proposed method combines components from consensus-based optimization algorithm with a newly introduced forcing term directed at the constraint set. A rigorous mean-field limit of the particle system is derived, and the convergence of the mean-field limit to the constrained minimizer is established. Additionally, we introduce a stable discretized algorithm and conduct various numerical experiments to demonstrate the performance of the proposed method.