Stationary probability density of stochastic search processes in global optimization

A method for the construction of approximate analytical expressions for the stationary marginal densities of general stochastic search processes is proposed. By the marginal densities, regions of the search space that with high probability contain the global optima can be readily defined. The density estimation procedure involves a controlled number of linear operations, with a computational cost per iteration that grows linearly with problem size.

💡 Research Summary

The paper introduces a novel analytical framework for estimating the stationary marginal probability densities of stochastic search processes used in global optimization. By modeling a generic stochastic optimizer as a continuous‑time Langevin dynamics, the authors derive the associated Fokker‑Planck equation that governs the evolution of the full joint probability density function (PDF) over the search space. Recognizing that solving the high‑dimensional Fokker‑Planck equation directly is infeasible, they focus on the marginal densities of each decision variable, which retain essential information about where the global optimum is likely to reside while dramatically reducing dimensionality.

The key technical contribution is a linear‑algebraic approximation scheme for the stationary marginal equations. Starting from the stationary condition (∂p/∂t = 0), the marginal Fokker‑Planck equation for a single coordinate xi is reduced to a one‑dimensional ordinary differential equation of the form

 d/dxi