A Mean-Field Optimal Control Formulation for Global Optimization

This paper is concerned with variational optimal control constructions whose solution yields a sampling algorithm. The particular form of the sampling algorithm considered here is a particle filter, designed to numerically approximate the solution to the global optimization problem. The theoretical significance of this study comes from its variational aspects. Specifically, the control input represents the solution of a mean-field-type optimal control problem. Its parametric counterpart, obtained when a parametric form of density is known a priori, is shown to be equivalent to the natural gradient algorithm. Explicit formulae for the filter are derived when the objective function is quadratic and the density is Gaussian. The optimal control construction of the particle filter is a significant departure from the classical importance sampling–resampling-based approaches.

[1]  Sean P. Meyn,et al.  Multivariable feedback particle filter , 2012, 2012 IEEE 51st IEEE Conference on Decision and Control (CDC).

[2]  Chi Zhang.  A particle system for global optimization , 2013, 52nd IEEE Conference on Decision and Control.

[3]  Fred Daum,et al.  Exact particle flow for nonlinear filters , 2010, Defense + Commercial Sensing.

[4]  Panos Stinis,et al.  Stochastic global optimization as a filtering problem , 2009, J. Comput. Phys..

[5]  Mark R. Morelande,et al.  Optimal parameterization of posterior densities using homotopy , 2011, 14th International Conference on Information Fusion.

[6]  R. Rubinstein.  The Cross-Entropy Method for Combinatorial and Continuous Optimization , 1999 .

[7]  A. Doucet,et al.  Gibbs flow for approximate transport with applications to Bayesian computation , 2015, Journal of the Royal Statistical Society: Series B (Statistical Methodology).

[8]  Yuhui Shi,et al.  Particle Filter Optimization: A Brief Introduction , 2016, ICSI.

[9]  Sean P. Meyn,et al.  Gain function approximation in the feedback particle filter , 2016, 2016 IEEE 55th Conference on Decision and Control (CDC).

[10]  Frederick E. Daum,et al.  Gromov's method for Bayesian stochastic particle flow: A simple exact formula for Q , 2016, 2016 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems (MFI).

[11]  M. Ledoux,et al.  Analysis and Geometry of Markov Diffusion Operators , 2013 .

[12]  G. Dullerud,et al.  A Course in Robust Control Theory: A Convex Approach , 2005 .

[13]  Sean P. Meyn,et al.  Synchronization of Coupled Oscillators is a Game , 2010, IEEE Transactions on Automatic Control.

[14]  Tryphon T. Georgiou,et al.  On the Relation Between Optimal Transport and Schrödinger Bridges: A Stochastic Control Viewpoint , 2014, J. Optim. Theory Appl..

[15]  R. Durrett.  Probability: Theory and Examples , 1993 .

[16]  Alain Bensoussan,et al.  The Master equation in mean field theory , 2014, 1404.4150.

[17]  Michael C. Fu,et al.  A Model Reference Adaptive Search Method for Global Optimization , 2007, Oper. Res..

[18]  Tryphon T. Georgiou,et al.  Optimal Steering of a Linear Stochastic System to a Final Probability Distribution, Part I , 2016, IEEE Transactions on Automatic Control.

[19]  Yongqiang Wang,et al.  Model-based Evolutionary Optimization , 2010, Proceedings of the 2010 Winter Simulation Conference.

[20]  Bin Liu,et al.  Posterior exploration based sequential Monte Carlo for global optimization , 2015, J. Glob. Optim..

[21]  Shun-ichi Amari,et al.  Natural Gradient Works Efficiently in Learning , 1998, Neural Computation.

[22]  Sean P. Meyn,et al.  Poisson's equation in nonlinear filtering , 2014, 53rd IEEE Conference on Decision and Control.

[23]  Stephen S.-T. Yau,et al.  Optimal control of the Liouville equation , 2007 .

[24]  Ping Hu,et al.  A Stochastic Approximation Framework for a Class of Randomized Optimization Algorithms , 2012, IEEE Transactions on Automatic Control.

[25]  M. Girolami,et al.  Riemann manifold Langevin and Hamiltonian Monte Carlo methods , 2011, Journal of the Royal Statistical Society: Series B (Statistical Methodology).

[26]  S. Reich.  A dynamical systems framework for intermittent data assimilation , 2011 .

[27]  Michael C. Fu,et al.  Particle Filtering Framework for a Class of Randomized Optimization Algorithms , 2014, IEEE Transactions on Automatic Control.

[28]  Sean P. Meyn,et al.  Feedback Particle Filter , 2013, IEEE Transactions on Automatic Control.