-
A Feasible Method for Constrained Derivative-Free Optimization
Authors:
Melody Qiming Xuan,
Jorge Nocedal
Abstract:
This paper explores a method for solving constrained optimization problems when the derivatives of the objective function are unavailable, while the derivatives of the constraints are known. We allow the objective and constraint function to be nonconvex. The method constructs a quadratic model of the objective function via interpolation and computes a step by minimizing this model subject to the o…
▽ More
This paper explores a method for solving constrained optimization problems when the derivatives of the objective function are unavailable, while the derivatives of the constraints are known. We allow the objective and constraint function to be nonconvex. The method constructs a quadratic model of the objective function via interpolation and computes a step by minimizing this model subject to the original constraints in the problem and a trust region constraint. The step computation requires the solution of a general nonlinear program, which is economically feasible when the constraints and their derivatives are very inexpensive to compute compared to the objective function. The paper includes a summary of numerical results that highlight the method's promising potential.
△ Less
Submitted 19 February, 2024;
originally announced February 2024.
-
Adaptive Finite-Difference Interval Estimation for Noisy Derivative-Free Optimization
Authors:
Hao-Jun Michael Shi,
Yuchen Xie,
Melody Qiming Xuan,
Jorge Nocedal
Abstract:
A common approach for minimizing a smooth nonlinear function is to employ finite-difference approximations to the gradient. While this can be easily performed when no error is present within the function evaluations, when the function is noisy, the optimal choice requires information about the noise level and higher-order derivatives of the function, which is often unavailable. Given the noise lev…
▽ More
A common approach for minimizing a smooth nonlinear function is to employ finite-difference approximations to the gradient. While this can be easily performed when no error is present within the function evaluations, when the function is noisy, the optimal choice requires information about the noise level and higher-order derivatives of the function, which is often unavailable. Given the noise level of the function, we propose a bisection search for finding a finite-difference interval for any finite-difference scheme that balances the truncation error, which arises from the error in the Taylor series approximation, and the measurement error, which results from noise in the function evaluation. Our procedure produces reliable estimates of the finite-difference interval at low cost without explicitly approximating higher-order derivatives. We show its numerical reliability and accuracy on a set of test problems. When combined with L-BFGS, we obtain a robust method for minimizing noisy black-box functions, as illustrated on a subset of unconstrained CUTEst problems with synthetically added noise.
△ Less
Submitted 22 March, 2022; v1 submitted 12 October, 2021;
originally announced October 2021.
-
On the Numerical Performance of Derivative-Free Optimization Methods Based on Finite-Difference Approximations
Authors:
Hao-Jun Michael Shi,
Melody Qiming Xuan,
Figen Oztoprak,
Jorge Nocedal
Abstract:
The goal of this paper is to investigate an approach for derivative-free optimization that has not received sufficient attention in the literature and is yet one of the simplest to implement and parallelize. It consists of computing gradients of a smoothed approximation of the objective function (and constraints), and employing them within established codes. These gradient approximations are calcu…
▽ More
The goal of this paper is to investigate an approach for derivative-free optimization that has not received sufficient attention in the literature and is yet one of the simplest to implement and parallelize. It consists of computing gradients of a smoothed approximation of the objective function (and constraints), and employing them within established codes. These gradient approximations are calculated by finite differences, with a differencing interval determined by the noise level in the functions and a bound on the second or third derivatives. It is assumed that noise level is known or can be estimated by means of difference tables or sampling. The use of finite differences has been largely dismissed in the derivative-free optimization literature as too expensive in terms of function evaluations and/or as impractical when the objective function contains noise. The test results presented in this paper suggest that such views should be re-examined and that the finite-difference approach has much to be recommended. The tests compared NEWUOA, DFO-LS and COBYLA against the finite-difference approach on three classes of problems: general unconstrained problems, nonlinear least squares, and general nonlinear programs with equality constraints.
△ Less
Submitted 19 February, 2021;
originally announced February 2021.
