Abstract:This paper proposes a constrained stochastic successive convex approximation (CSSCA) algorithm to find a stationary point for a general non-convex stochastic optimization problem, whose objective and constraint functions are non-convex and involve expectations over random states. Most existing methods for non-convex stochastic optimization, such as the stochastic (average) gradient and stochastic majorization-minimization, only consider minimizing a stochastic non-convex objective over a deterministic convex set. The proposed CSSCA algorithm can also handle stochastic non-convex constraints in optimization problems, and it opens the way to solving more challenging optimization problems that occur in many applications. The algorithm is based on solving a sequence of convex objective/feasibility optimization problems obtained by replacing the objective/constraint functions in the original problems with some convex surrogate functions. The CSSCA algorithm allows a wide class of surrogate functions and thus provides many freedoms to design good surrogate functions for specific applications. Moreover, it also facilitates parallel implementation for solving large-scale stochastic optimization problems, which arise naturally in today's signal processing such as machine learning and big data analysis. We establish the convergence of the CSSCA algorithm with a feasible initial point, and customize the algorithmic framework to solve several important application problems. Simulations show that the CSSCA algorithm can achieve superior performance over existing solutions.

Chance-constrained Quasi-Convex Optimization with Application to Data-Driven Switched Systems Control

Time-delayed Feedback Control Optimization for Quasi Linear Systems under Random Excitations

Stochastic Optimal Control of Quasi Non-Integrable Hamiltonian Systems with Stochastic Maximum Principle

Stochastic Optimal Control for the Response of Quasi Non-Integrable Hamiltonian Systems

Constrained Optimal Control of Switched Systems and Its Application

Data-driven Control of unknown Switched Linear Systems using Scenario Optimization

Chance Constrained Stochastic Optimal Control Based on Sample Statistics With Almost Surely Probabilistic Guarantees

Stochastic Successive Convex Approximation for Non-Convex Constrained Stochastic Optimization

Accelerated Stochastic Optimization Methods under Quasar-convexity

ADVANCES IN RESEARCH ON NONLINEAR STOCHASTIC OPTIMAL CONTROL OF QUASI-HAMILTONIAN SYSTEMS

Chance-constrained optimization for contact-rich systems using mixed integer programming

Chance Constrained Stochastic Optimal Control for Linear Systems with a Time Varying Random Control Matrix

Switching Time Optimization for Binary Quantum Optimal Control

An Empirical Quantile Estimation Approach to Nonlinear Optimization Problems with Chance Constraints

Parabolic Optimal Control Problems with Combinatorial Switching Constraints, Part I: Convex Relaxations

An Empirical Quantile Estimation Approach for Chance-Constrained Nonlinear Optimization Problems

Stochastic linear-quadratic optimal control problems with non-Markovian regime switching

Uncertain standard quadratic optimization under distributional assumptions: a chance-constrained epigraphic approach

Binary Quantum Control Optimization with Uncertain Hamiltonians

Optimal Control of Constrained Stochastic Linear-Quadratic Model with Applications

Stochastic linear quadratic optimal control problems with regime-switching jumps in infinite horizon